Aug 2026 · 5 min read
Why 70 to 95% of enterprise AI projects fail, and why it is not the technology. Risk, return on investment and bias must be tackled together.
Brent C. Hector, MCGI, PMP®
Strategic programme and enterprise delivery leader with 20+ years of senior leadership across public administration, enterprise technology, financial services and humanitarian operations. Board Director at iGovTT.
MCGI · PMP® · PSM I · PSPO I · PAL-EBM · Harvard Humanitarian Academy
About
Strategic programme and enterprise delivery leader with 20+ years of senior leadership across public administration, enterprise technology, financial services and humanitarian operations. I led a government project unit responsible for people, annual budgeting, procurement, financial controls and a portfolio of programmes and initiatives. I have managed complex technology delivery in a regulated banking environment across three Caribbean jurisdictions, and I contribute to board-level governance at Trinidad and Tobago’s national ICT company. I combine strategic planning and operational management with practical programme controls, stakeholder leadership and delivery oversight.
I’m based in Trinidad and Tobago. I own outcomes, not just outputs, and I keep decision logs and risk registers audit-ready. I also lead disaster relief operations through ITNAC.
How I work
Use one, or all three.
Know what to do before you spend. I help boards and leaders decide what to build, in what order, and at what cost.
In practice: at iGovTT I review strategic plans, annual operating plans and draft budget estimates with the CEO and Board.
Turn the plan into something people can run: structures, roadmaps, governance and reporting.
In practice: with ITNAC I built KoboToolbox assessment tools that produce datasets aligned to UN and Red Cross reporting.
Land it on time, on budget and audit-ready, with vendors and stakeholders aligned.
In practice: as a project manager in the PMO at Massy Technologies, Agile governance helped cut programme timelines and costs by approximately 25%.
How I think
Six stories, six endings. Tap a card to see how each one ends.
Once upon a time, a programme had a growing backlog of engineering changes.
Every day, reviewers worked through approvals.
One day, management added review capacity, and approvals sped up.
Because of that, construction raced ahead and released a fresh wave of change requests.
Because of that, the backlog came back, and management asked for even more reviewers.
Once upon a time, a transformation programme fell behind.
Every day, specialists did the delivery work.
One day, management asked for more progress reporting.
Because of that, specialists lost time to reports, and delivery slowed.
Because of that, the new delays triggered still more reporting.
Once upon a time, two delivery partners shared one integration project.
Every day, they traded evidence about a recurring failure.
One day, each blamed the other: software configuration or operating procedures.
Because of that, they ran separate investigations, each with credible but incomplete evidence.
Because of that, mistrust grew as the deadline came closer.
Once upon a time, a programme combined a stable facility build with an experimental digital service.
Every day, both teams pressed toward shared regulatory milestones.
One day, leadership chose one method for everything.
Because of that, the build lost its controls, or the service lost its room to learn.
Because of that, the shared interfaces began to slip.
Once upon a time, a global technology programme ran smoothly across regions.
Every day, regional teams made local calls.
One day, a sudden regulatory change landed.
Because of that, teams waited for central approval, and responses slowed.
Because of that, some regions acted alone, and compliance exposure grew.
Once upon a time, a product programme planned once a quarter.
Every day, customer needs shifted a little more.
One day, engineering, compliance and service teams collided over priorities.
Because of that, the same conflicts returned every planning cycle.
Because of that, the sponsor asked to plan faster, but the regulatory gates could not move.
Your programme is the next scene.
Where I work
Public-sector ICT governance and programme delivery in Trinidad and Tobago, from the Ministry of Education to the iGovTT board.
Enterprise technology delivery for regional financial institutions and corporations, across two separate engagements:
Relief logistics, MEAL assessment and donor-compatible reporting for Caribbean disaster responses, including Hurricane Maria in Dominica and Hurricane Matthew in Haiti.
Experience
Credentials
Writing
22 articles on delivering change in government and business. Start with the three marked, or filter by theme. Each one opens right here.
Showing 22 articles
Aug 2026 · 5 min read
Why 70 to 95% of enterprise AI projects fail, and why it is not the technology. Risk, return on investment and bias must be tackled together.
Jul 2026 · 4 min read
Failure starts before the plan, when nobody agrees what problem they are solving. A story from 123 school boards, and how to plan constantly but never completely.
May 2026 · 6 min read
A practitioner’s guide to what blocks agile in Caribbean ministries: the comfort of the paper trail, agile theatre and weak leadership, and how to start small but properly.
Nov 2026 · 3 min read
When operations break, boards reach for software. Why first-principles thinking and fixing the process must come before automation.
Nov 2026 · 5 min read
A framework will not fix poor leadership. On the danger of becoming a slave to frameworks, and the skill of challenging your own thinking.
Nov 2026 · 5 min read
Most Caribbean businesses have a digital plan, but far fewer deliver it on time. Where projects stall and what structured delivery changes.
Sep 2026 · 2 min read
What if you are paying people to wait? On over-processing as waste, and delivering the required outcome to the required standard without wasting resources.
Sep 2026 · 5 min read
We promote technically capable people into leadership and assume they can lead. Why psychological safety and real development matter more as AI accelerates.
Sep 2026 · 5 min read
How moving from project management to Evidence-Based Management shifts leaders from output to outcome, and why the hardest part is the leadership mindset.
Sep 2026 · 5 min read
Disaster resilience as the core function of Caribbean sovereignty: three failure modes (physical, fiscal and digital) and a policy test.
Aug 2026 · 4 min read
A Star Trek take on why every public sector programme hears “no” before “yes”: six engineering objections decoded for programme leaders.
Aug 2026 · 4 min read
A personal log of joining the Ministry of Public Administration in 2006, where my theories about slow government met a first assignment with TCS, told in a Star Trek voice.
Jul 2026 · 7 min read
A sharper take from the CTU AI forum: sharing a vision is passive, governance is the main text, and a plan needs room for being wrong.
Jul 2026 · 4 min read
Reflections from the first CTU Caribbean AI Forum: national AI success depends on governance, investment and delivery, not technology alone.
Jul 2026 · 3 min read
What the Agile Manifesto’s pioneers teach about uncertainty: make change cheap, speculate instead of over-planning, rely on hard truths and decentralise decisions.
Jul 2026 · 3 min read
Scrum is not Agile, but calling it just the method rewrites history. A pushback on flattening Scrum into procedure.
Jun 2026 · 2 min read
Beyond the triple constraint: a reflection on why time is the one constraint you cannot buy more of.
Jun 2026 · 5 min read
“Digital government” is a slogan until a citizen completes a transaction. Lessons from the UK Government Digital Service, the portal trap, and measuring performance.
May 2026 · 2 min read
The best coordinators often have no line authority. How to deliver across security, compliance, operations and technology teams without budget or reporting lines.
Mar 2026 · 4 min read
Why government IT launches become ceremonies instead of change: structural incentives, institutional culture and the case for organisational change management.
Jan 2025 · 2 min read
Working as an Agile practitioner inside phase-gated environments: where the two approaches clash, and how to measure impact and overcome resistance.
Jul 2022 · 1 min read
Is agility just new vocabulary and ceremonies, or a deeper philosophical shift? A short reflection on stable teams and servant leadership.
Free workbook
A practical workbook from my Working in Projects framework, built for people who deliver real programmes. Free. Enter your name and email and I’ll send you a copy.
Your details are used only to send the workbook and to reply to you.
Process
Four steps, no surprises.
Tell me your priority and the support you need. Please leave out confidential details.
We cover context, stakeholders, constraints and the type of support that fits.
If there is a fit, I define the work, roles, deliverables, timing and terms.
Work starts once scope and arrangements are agreed.
ERP rollout, digital government initiative or an operation under pressure. Let’s talk about what it needs.
LinkedIn · Trinidad and Tobago
A practitioner's guide to what actually blocks transformation in the Caribbean public sector — and what to do about it.
The first time agile enters a ministry, the room tells you everything.
Technical staff are curious, even relieved. Finally, a way to show progress in small pieces instead of vanishing for months behind a plan nobody believes.
Middle managers are cautious. They want to know who owns the backlog and whether this makes their lives easier or riskier.
The tension sits higher up the table. It appears when senior leaders hear what agile quietly asks of them:
We do not fully know what we are building yet.
For a public-sector leader, that sentence feels dangerous. The discomfort is understandable. It is also the main barrier.
It surfaces the moment senior leaders understand what agile is quietly asking them to accept:
We do not fully know what we are building yet.
For many public-sector leaders, that sentence feels dangerous. And that discomfort - entirely understandable - is precisely the barrier that makes agile transformation in government hard.
Government loves a paper trail, and for good reason. Every dollar must be justified. Every procurement can be challenged. Every decision may one day be explained to an auditor, a minister or the press.
So the system protects itself with business cases, fixed scope, signed approvals and steering committees. These controls are necessary. I am not arguing against them.
The trouble starts when the paper trail matters more than the truth on the ground. A project can look well governed and still be failing.
The dashboard is green. The minutes are clean. Then, the week before go-live, everyone learns what users suspected and the technical team already knew: the thing being delivered is not the thing people need.
Traditional delivery checks accountability at the end. Agile checks it every sprint. What did we deliver? What did users tell us? What changed because of it? What decision is blocking us?
You can soften a status report. You can delay bad news through governance layers. You cannot easily hide a sprint review where nothing useful was produced.
Government wants certainty before delivery starts. Agile builds certainty through delivery. That shift is uncomfortable, and necessary.
Agile becomes real when a leader lets a team act on what it has learned. Not permission to skip governance. Permission to make evidence-based decisions within clear boundaries.
Most organisations miss this. They train Scrum Masters, build backlogs and run stand-ups. But the real question is not whether the team uses agile language. It is who may decide when new information appears.
In many ministries, decisions still crawl upward. The team finds a problem. The project manager escalates. The steering committee asks for more information. The next meeting is scheduled. The team waits.
By the time the answer returns, the situation has changed. That is how agile dies in a bureaucracy. Nobody rejects it. It dies waiting for approval.
Decision rights, meaning who decides what, at which level, on what evidence, are not an administrative detail. They decide whether agile stands or collapses.
Some agile looks impressive from the outside. Daily stand-ups. Walls of sticky notes. A tidy backlog, a good Scrum Master, a velocity chart, steering reports full of agile terms.
Underneath, nothing has changed. The team cannot adjust scope without escalation. Users see the product late. Progress is measured by activity, not usable value.
That is agile theatre. The organisation has taken the rituals and kept the instincts of waterfall. It wants speed with the comfort of old controls, and user feedback only after the big decisions are made.
That contradiction always shows up in delivery.
Paper trails delay bad news. The retrospective does the opposite. Leaders often see it as another meeting. It is where a team asks what slowed it down, what it misunderstood and what to change before the next sprint.
In government this is quietly radical. Bureaucracies are built to defend decisions already made, not to say "We got that wrong, and here is what we are changing."
Public institutions work under scrutiny, so the instinct makes sense. But hiding small failures produces larger ones later. A good retrospective strengthens accountability by correcting course while it is still cheap.
That is what mature agile gives government. Control, with the truth arriving earlier.
We cannot import a model from the UK, Australia or a large technology firm and expect it to work unchanged. Our context is different.
Many agencies have small teams and few specialists. The same people sit on several committees. Procurement is slow. Decisions are split across ministries, boards, vendors and partners.
Agile assumes a product owner who can decide, users who can be reached, a team with technical, business, design and testing skills, fast removal of blockers, and leaders who let evidence change the plan.
In many Caribbean agencies, those conditions do not exist yet. That does not mean agile cannot work here. It means adopting the vocabulary is not enough.
The honest question is not "Are we doing agile?" It is "Have we created the conditions for agile to produce value?" If not, the work starts there.
Ask one question of any agile project: when does the citizen first respond to what is being built? The citizen might be a business owner, student, patient or public officer.
If the answer is "at launch", the agile is decorative.
A licensing system businesses cannot navigate. A health platform nurses find impractical. A student portal that creates more support calls than it solves. Each one digitised the confusion.
These failures rarely come from carelessness. Users arrived too late. Real agile brings them in while there is still time to change the outcome.
Do not begin with a grand mandate. Begin with one visible, well-chosen project.
Skip the symbolic pilot, the rebranded project and the donor experiment that ends in a final report. Pick one with a clear user group, manageable scope, an engaged sponsor and a team that can act on what it learns.
The real goal is to teach the institution a new way of working. A good first project gives leaders something to see, users something to react to and the team something to improve.
Transformation takes root through visible delivery.
Major public-sector technology failures are rarely just bad project management. The institution failed to learn fast enough. One department knew one part, a vendor another. Nobody had the full picture in time.
For government, agile is a leadership discipline.
Agile does not ask senior leaders to give up control. It asks for a better form of it: control through clarity, priorities, evidence, short feedback loops, fast decisions and visible value.
The old model says: come back when everything is defined. The agile model says: show me what we know, what we have learned, what users are saying and what decision you need from me.
That posture takes humility, courage and discipline. The hard part is not teaching teams to work in sprints. It is teaching institutions to learn before they fail.
A ministry can run stand-ups and still be waterfall. An agency can keep a backlog and still avoid hard decisions. A programme can report velocity and deliver little.
The region's future will not be decided by who adopts agile language fastest. It will be decided by who changes how decisions are made.
Agile does not fail in government. Governance fails agile.
The FBI is one of the most entrenched bureaucracies in the United States. After a decade of failed waterfall attempts and hundreds of millions of dollars spent, its case management system, Sentinel, was still in trouble.
Sentinel was meant to connect the intelligence that siloed, paper-based systems failed to join before 11 September 2001. It was a year behind schedule, and new defects arrived faster than old ones were fixed.
A lean team took it back in-house, applied agile principles and delivered. Jeff Sutherland, Scrum co-founder, documents the turnaround in Scrum: The Art of Doing Twice the Work in Half the Time.
Neither better technology nor more funding saved it. A leadership decision did. Leaders chose to treat uncertainty as a condition of the work, not a management failure.
That is what agile asks of Caribbean leaders. Less a methodology than a disposition.
The conditions that nearly broke Sentinel exist in some form in every government in this region. Clearing them takes more than a framework. It takes someone who has sat in the room where these decisions are made.
If you are ready to have that conversation, I am ready to lead it. Get in touch and tell me where your first project would start..brenthector.com
Writing
(Captain's Personal Log // Stardate 0603.1. Planet: Public Service.)
Until 2006, government was something I watched from the outside.
Like plenty of people in the private sector, I had strong opinions. Government moved too slowly. Everything seemed buried in paperwork. Meetings produced more meetings. I was convinced I knew why projects struggled.
Then I joined.
My theories lasted about a week.
I had barely found my desk in the ICT Division of the Ministry of Public Administration when I was assigned to what felt like my first away mission. Work with Tata Consultancy Services (TCS) on the post implementation review of the Government's largest ICT programme at the time: Project IHRIS, the Integrated Human Resource Information System, together with the Integrated Global Payroll (IGP).
I was the new guy. I was still learning people's names. Suddenly I was sitting in rooms with senior public servants, project managers, technical teams, HR professionals and international consultants, trying to answer a question nobody seemed comfortable asking.
A project had cost more than $200 million
So why wasn't anyone getting the value they expected?
Like any software developer, I assumed we'd spend our days hunting bugs, bad code or failed integrations.
We didn't.
The software wasn't the biggest problem.
That surprised me then. It doesn't anymore.
On paper, the project looked like a success.
The software had been delivered. Training was completed. User Acceptance Testing had been signed off. The contracts were finished. Every project report would probably have shown green.
If you stopped there, you would have called it a success.
Then the interviews started.
We met with the implementation team, technical staff, HR officers, payroll officers, department heads, power users and everyday users across the public service.
A different story emerged.
The system worked.
People simply weren't using it.
That wasn't because people were resisting technology for the sake of it. The public service HR landscape was, and still is, fragmented. The Personnel Department owned part of the process. PMCD owned another. The Public Service Transformation Division had responsibilities. The Service Commissions Department had others. Payroll belonged to the Comptroller of Accounts under the Ministry of Finance.
One employee's journey crossed several organizations.
The software connected the process.
The organizations never really did.
Payroll survived because salaries have to be paid. Funny how that creates urgency.
Almost everything else struggled.
Benefits administration. Position management. Recruitment. Pension management. Workforce planning. Leave administration.
Most of those modules never became part of everyday work.
What struck me most wasn't what people were doing.
It was what they kept doing.
Departments that were supposed to own their parts of the system kept sending requests back to the project office.
Need to update an employee record?
Call the project office.
Need information changed before payroll?
Call the project office.
Need a report?
Call the project office.
Somehow, the implementation team had quietly become the operations team.
Nobody announced it.
Nobody designed it that way.
It just happened.
Then we found something even harder to ignore.
Many of the people who had signed off on User Acceptance Testing couldn't clearly explain what they had approved.
The signatures were real.
The ownership wasn't.
The more I listened, the more I realised something was missing from the very beginning.
Nobody had built a shared understanding of the business problem the programme was solving.
There was no agreed picture of what success looked like after go live. No clear benefits assigned to business owners. No one measuring whether those benefits were ever realised.
The technology had been implemented.
The organisation hadn't changed.
One visit to Tobago brought all of that into sharp focus.
We weren't investigating software failures.
The software was doing exactly what it had been built to do.
Meanwhile, public servants were still waiting months, sometimes years, for pension and gratuity payments. HR officers were still digging through paper files. Salary adjustments still crawled through manual processes.
The irony was hard to miss.
Many of those problems already had a solution sitting on the server.
The review eventually produced more than fifty recommendations. Governance. Business ownership. Change management. Operational transition. Benefits realisation. They were practical recommendations. None of them required new technology.
Then came one final lesson.
The person who chaired the original implementation steering committee also chaired the post implementation review.
That never sat comfortably with me.
A post implementation review works best when it can challenge assumptions without looking over its shoulder.
Many of the recommendations faded away.
The report existed.
The lessons didn't spread.
Nearly twenty years have passed since those meetings.
Technology has changed beyond anything we imagined back then. Today the conversation is about cloud, AI and digital government.
What hasn't changed nearly as much are the conversations.
I still hear organisations asking which system they should buy.
I hear less discussion about who will own it after the project ends.
Who will change the business processes.
Who will be accountable for the benefits.
Back in March 2006, I thought I'd joined an IT department.
Looking back, I had actually enrolled in a masterclass on organisational change.
Project IHRIS taught me something I have carried into every programme since.
Technology is usually the easy part.
Changing how an organisation thinks, works and takes ownership is where the real work begins.
(Captain's Personal Log.
Computer, archive this one.
I have a feeling we'll be reading it again.)
Most Caribbean businesses now have a digital plan. Fewer deliver it on time.
Digicel Business surveyed more than 2,500 regional organisations in 2024. It found that 77% had a digital strategy, up from 61% in 2022. It also found that 60% of those surveyed were behind schedule on the rollout.
The gap between plan and delivery is a project management problem. This article looks at where Caribbean projects stall, what recent data says about fixing it, and why it matters for firms that want regional and international customers.
I use “digital trust” to mean a buyer’s confidence that your firm will deliver what it promises, on systems that work. It comes from finished projects, not from a good website.
A logistics firm in Port of Spain can bid for the same regional contract as a firm in Panama or Miami. The buyer may never visit either office. The question they ask is simple: can this firm deliver?
Evidence answers that question. Systems that went live, dates that held and data that stayed secure all count. A polished website does not.
The Caribbean Development Bank estimates that micro, small and medium enterprises make up 70% to 85% of businesses in its borrowing member countries. They contribute 60% to 70% of GDP and about 50% of employment.
Small firms have small teams. A digital project pulls in the same people who run daily operations, so a delay costs them more. The Digicel survey points the same way: 34% of respondents said key resources were consumed by operational issues.
The top three barriers were systems not properly integrated (35%), key resources consumed by operational issues (34%), and inadequate priority setting and focus from senior leadership (31%).
Two of the three are about people and priorities, not technology. Digicel sells connectivity and digital services, so treat the survey as one input and not a verdict. It is still the largest regional dataset I found for 2024.
An IDB report on six Caribbean economies (The Bahamas, Barbados, Guyana, Jamaica, Suriname, and Trinidad and Tobago) found that productivity is 26% to 35% higher in innovative firms than in non-innovative ones. The figures compare firms. They do not prove that innovation caused the gains.
The same report lists barriers: regulatory hurdles, limited access to finance, and skills mismatches in the labour market. IDB’s Caribbean General Manager Anton Edmunds said innovation is “one of the key drivers for growth and resilience in the Caribbean.”
PMI’s 2026 Pulse of the Profession found that 81% of project professionals say projects have become more complex. Teams that manage complexity well reported an 88% success rate. Teams that managed it poorly reported 14%.
Projects that used a framework succeeded 72% of the time, against 61% for those that did not. Yet 35% of professionals used no framework on their latest complex project. These are survey correlations, and PMI is a professional body that promotes frameworks. The direction is still consistent with what I have seen.
As a project manager at Massy Technologies and PBS Technology Group, I managed Oracle EBS, IBM Maximo, Power BI and infrastructure workstreams. Agile governance across them cut programme timelines and costs by about 25%. I also trained 17 IT Operations staff in Agile and Scrum and wrote Scheduling and Work Breakdown Structure learning modules for use across the organisation. My banking technology work included a SailPoint identity and access management rollout for 3,500 users.
These are results from my own projects, not a regional average. Training mattered as much as the method. A method only stays in place if the people running the next project understand it.
A shared schedule, a common work breakdown structure and a regular status report show a customer, a board or a lender where a project stands. A small team can produce the same records as a large one. That is the part of the “level playing field” argument that holds.
The Digicel barriers are not fixed by buying software. Each one maps to a management decision:
Unintegrated systems (35%): map how new systems connect to existing ones before purchase, not after. Resources consumed by operations (34%): assign named delivery time to project staff and protect it. Weak senior focus (31%): name one executive owner for each project, with the authority to settle priority conflicts.
Tools support these decisions. They do not make them.
Public bodies and private firms in the Caribbean increasingly depend on each other’s systems. That only works if both sides finish what they start.
Marlon Cooper, CEO of Symptai, described a digitally mature Caribbean this way: “Doing business is easy, because almost all services are online.” Each of those services exists because someone delivered a project.
It is a buyer’s confidence that a firm will deliver what it promises on systems that work. This article uses the term in that practical sense. It is not a formal standard.
The 2024 Digicel survey named three main barriers: systems not integrated (35%), staff consumed by operational issues (34%), and weak priority setting from senior leadership (31%). Sixty per cent of respondents reported delays.
The Caribbean Development Bank estimates that micro, small and medium enterprises are 70% to 85% of businesses and contribute 60% to 70% of GDP. Other published estimates differ, so check the source and year.
PMI’s 2026 research found a 72% success rate with a framework and 61% without. This is a correlation. In my own programmes, Agile governance cut timelines and costs by about 25%.
No. The leading barriers in the Digicel survey involve integration, staff capacity and leadership focus. Tools help with visibility and consistency, but people have to make those decisions.
Name one senior owner for each digital project, protect time for the people delivering it, and map system connections before buying. Each step answers one of the three barriers above.
Brent C. Hector is a Strategic Programme Leader, Technology and digital transformation consultant who helps Caribbean businesses and government build structured delivery capabilities. He has led enterprise software and identity infrastructure workstreams across the Caribbean, including policy development, Oracle EBS, IBM Maximo, Power BI, and SailPoint implementations.
If your digital strategy is falling behind schedule, the fix rarely starts with buying more software. It starts with setting clear priorities, protecting delivery time for your team, and putting a practical framework in place.
To learn more about modernising your delivery or to discuss your next project, visit brenthector.com.
“Digital government” is one of those phrases that sounds impressive in a strategy document.
It is broad enough for ministers to announce. Attractive enough for donors to fund. Flexible enough for consultants to package. But ask two senior officials in the same ministry what it actually means, and you may hear two very different answers.
One may speak about infrastructure. Another may speak about citizen experience. A third may speak about portals, cybersecurity, data, legislation, or payment systems.
None of them may be wrong. But are they speaking about the same thing?
That is where the problem begins.
The ambiguity around “digital government” is not merely a language issue. It has real operational consequences. Programmes fail because the people funding them, designing them, managing them, and governing them are often optimising for different outcomes without realising it.
After years of working around public-sector transformation, I have found one framing far more useful:
Digital government is not a destination. It is a capability built service by service, transaction by transaction.
There is no magical moment when a government “becomes digital.” What exists are individual services — a business registration, a passport application, a tax return, a health record, a social support application — that either meet the standard of a good digital service, or they do not.
That distinction matters.
Instead of asking, “How digitally transformed is our government?” — a question that usually produces vague answers — leaders should be asking something much more practical:
For the twenty services citizens use most often, what percentage of people can complete them digitally, end to end, without help?
That question can be measured. It can be tracked. It can be assigned. It can be improved.
And that is where real transformation begins.
When the UK Government Digital Service launched in 2011, it did not begin with a portal strategy or a branding exercise. It began with a simple but powerful premise: government services should meet the same basic usability standard that people had come to expect from good consumer digital products.
Not the same visual design. Not the same commercial model. The same fundamental principle.
A person should be able to complete what they came to do without needing help.
That is a radical idea in government.
The work that followed was not abstract “digital transformation.” It was disciplined service redesign. One service at a time, teams identified where citizens were failing, studied why they were failing, and removed the specific barriers preventing completion.
That is the part many governments miss.
The transformation was not in the website. It was in the service.
The GDS Service Standard created a clear definition of what a good digital government service should look like. It did not simply ask, “Is the technology working?” It asked deeper questions:
Can the user complete the service? Does it work for people with common access needs? Is the service simple, inclusive, measurable, and continuously improved?
Those are the questions that separate technology implementation from public value.
The results were measurable. HMRC’s online self-assessment saw significant growth in digital completion. The Lasting Power of Attorney service reduced average processing time from weeks to days. But these gains did not happen because someone announced a digital transformation strategy.
They happened because specific services were redesigned, measured, governed, and improved.
That is the lesson.
One of the most common mistakes in digital government is confusing a portal with transformation.
A portal can be useful. It can improve navigation. It can create a consistent front door for citizens. It can make government feel more coherent from the outside.
But a portal does not transform the underlying service.
If a citizen visits a modern government website, clicks on a beautifully designed link, downloads a PDF form, prints it, signs it in wet ink, scans it, emails it, and then still has to visit an office in person, what has actually changed?
The journey to the bad experience has improved. The bad experience itself remains intact.
That is not transformation. That is digital decoration.
A portal is the visible tip of a much larger transformation iceberg. The real work sits below the surface: process redesign, data integration, legal authority for electronic submissions, backend workflow, staff adoption, records management, identity verification, payment integration, and service accountability.
This work is not glamorous. It rarely produces a ribbon-cutting moment. It does not always photograph well.
But it is the work that changes the citizen’s experience.
Judging digital transformation by the quality of a portal is like judging a hospital’s medical capability by the design of its reception desk. The reception matters, but it is not the treatment.
The discipline that separates genuine digital progress from performance is measurement.
Not press releases. Not launch events. Not the number of services “placed online.”
The real questions are:
What percentage of users who start the service complete it?
How satisfied are they with the experience?
How many transactions are now fully digital compared with paper or in-person channels? Where are users dropping off?
Why are they dropping off?
These metrics are unforgiving, but they are necessary.
A service with a 30% digital completion rate is not yet a successful digital service. It is a digital obstacle with a paper fallback. And that fallback should not be treated as comfort. It is evidence that the digital service has not yet become good enough to replace the old process.
For Caribbean governments, this is where the investment conversation needs to mature.
The priority is not always another portal. It is measurement infrastructure: the analytics, reporting, governance, and service ownership needed to know where citizens are failing and why.
Without that data, leaders are managing perception, not performance.
Every time a citizen successfully completes a government service online, their confidence in the digital channel increases. They are more likely to try the next service. More likely to recommend it. More likely to believe that government can, in fact, work differently.
But every failed attempt has a cost.
A broken link. A timeout error. A form that cannot be submitted. A request to “come in person” after everything was supposedly completed online.
Each failure teaches the citizen a lesson: do not trust the digital channel.
And once that trust is lost, the behavioural cost of getting people to try again becomes much higher.
Trust in digital government is not built by slogans. It is built one completed transaction at a time
No communications campaign can compensate for a service that does not work. No portal redesign can substitute for a transaction that cannot be completed. No ministerial announcement can overcome repeated citizen frustration.
The question after any digital service launch should not be, “How many people visited the site?”
The better question is:
How many people completed what they came to do?
That is the real measure of digital government.
Because in the end, there is no digital government in the abstract.
There are only services.
And each one either builds trust, or erodes it.
Navigating unpredictability has become the defining challenge of leadership today. When volatility hits, the instinct is often to tighten controls, demand more detailed forecasts, and build heavier plans.
But the creators of the Agile Manifesto had a different perspective.
To the seventeen pioneers who met in Utah twenty-five years ago, uncertainty was not a problem to be solved. They viewed it as an inescapable property of any complex system. If you want to build a truly resilient organization, you have to stop trying to plan away the unknown. You need to design your business to thrive inside it.
Here is how the original signatories tell us to handle uncertainty.
Traditional planning assumes we can predict the future if we just analyze the data hard enough. This mindset leads to massive upfront requirements, rigid roadmaps, and detailed contracts. The signatories argued that this approach is fundamentally flawed.
Instead of wasting time and energy trying to make the future certain, focus on making the cost of changing your mind cheap.
Kent Beck, the creator of Extreme Programming, used a driving analogy to explain this. You do not point a car in a direction, lock the steering wheel, and step on the gas. That is what rigid upfront planning does. Instead, you make constant micro-corrections based on the road ahead. To do this safely at high speeds, you need highly responsive steering.
In business, that steering translates to loose coupling, flexible architecture, and processes designed to pivot on a dime. If a decision is incredibly difficult to undo, it is a high-risk decision. Build reversibility into your strategy.
When you are operating in deep uncertainty, you cannot analyze your way to a solution. You have to act your way into it.
Jim Highsmith, author of *Adaptive Software Development*, intentionally chose the word "speculate" over "plan." Planning implies a high degree of certainty and control. Speculating acknowledges that we are making an educated guess that must be validated through real-world execution.
Alistair Cockburn championed the use of small, functional "probes." When faced with an unknown technical or market path, do not write a long strategy document. Deliver a tiny, working slice of your idea to see how the market actually reacts. Run quick, low-cost experiments designed solely to gather information.
Status reports, green-lit milestone charts, and subjective progress reviews can easily mask underlying risks. The pioneers believed that the only real measure of progress is a working product in the hands of a user.
Put your ideas in front of actual customers as quickly as possible. When a real user interacts with what you have built, the uncertainty evaporates. They will either use it, or they won't. They will find value in it, or they will ask for something else. That feedback is the only truth mechanism that matters.
The most dangerous variable in a volatile market is the time that passes between making an assumption and testing it. If you are going to make a mistake, you want to find out immediately.
Jeff Sutherland and Ken Schwaber, the co-creators of Scrum, designed the framework around empirical process control. If you do not know where the market or the technology is going, a long planning cycle is a liability.
Keep your feedback loops incredibly short. This limits your financial and operational exposure to just a few days of wrong assumptions. Shrinking your cycle time is the single best risk-management tool you have.
Centralized command-and-control structures fall apart in fast-moving environments. Information takes too long to travel up the hierarchy, and decisions take too long to filter back down.
The people closest to the work are always in the best position to make decisions when unexpected variables arise. Leaders should provide clear intent and boundaries, and then trust autonomous, cross-functional teams to figure out the execution.
Agile was never meant to be a set of rigid ceremonies, certified checklists, or dogmatic rules. It was designed as a survival strategy for creative professionals operating in an unpredictable world. Meeting uncertainty with adaptability, rather than resistance, is how we build organizations that last.
*What strategies is your leadership team using to shorten feedback loops and reduce the cost of change this quarter? Let's discuss in the comments.*
How shifting from project management to Evidence-Based Management changes the questions leaders ask, the choices they make, and the outcomes they deliver.
Back in 2021, I wrapped up the Professional Agile Leadership – Evidence-Based Management (PAL-EBM) certification through Scrum.org.
At the time, I saw it simply as a practical step forward in building my Agile leadership skills.
Five years on, my perspective has shifted entirely.
That experience fundamentally changed how I view projects, products, technology, and capital investment. But honestly, the biggest shift wasn't in tools or frameworks—it was in the questions I started asking.
Before this shift, most project and technology management conversations revolved strictly around delivery mechanics:
What are the requirements?
What’s the scope?
What’s the schedule?
What’s the budget?
When do we launch?
All of these are still important. They are just no longer enough.
Now, I consistently bring teams back to a much simpler question:
"What value are we trying to create, for whom, and how will we know when we’ve actually created it?"
It sounds obvious on paper. In practice? It is far from it.
You can deliver a project with total precision and still completely miss the mark.
The system goes live.
The scope checklist is completed.
The project closes on time and within budget.
...and yet, you still fail to solve the core business problem.
That is the distinction between output and outcome:
Output is what we make.
Outcome is what changes because of what we made.
EBM locked this distinction into my thinking.
Scrum.org describes Evidence-Based Management as an approach centered on goals, measures, hypotheses, experiments, and continuous learning—all designed to improve customer outcomes, organizational capability, and business results.
That framework reframed how I evaluate investments. Delivery alone isn't proof of success—it's just proof that something was built. Proving value is a different level altogether.
As I moved away from old-school project delivery thinking, a new set of questions took priority:
Who is the user? Who is the buyer?
What problem are we actually solving, and how urgent is it?
Where does the real value sit?
What evidence shows people actually want this?
Will anyone use it? Are they willing to pay for it?
How much will it cost to build and maintain long-term?
And perhaps the hardest question of all:
"Should we build this at all?"
That question frequently gets lost. Technology teams are exceptionally good at building things. But just because you can build something doesn't mean you should.
This is why user-centered design and design thinking are indispensable:
Understand the user.
Unpack the real problem.
Test your assumptions early.
Learn from actual market evidence.
Then decide if it’s worth investing in.
Whether you are building a commercial software product or a government service, this logic applies. A ministry may fund a service, an agency may run it, citizens use it, and taxpayers foot the bill—the core value proposition still has to hold up.
In EBM, the four Key Value Areas (KVAs) become the primary strategic lenses:
Current Value: What value are we creating right now?
Unrealized Value: What potential value is sitting there uncaptured?
Time-to-Market: How quickly can we get something in front of users to start learning?
Ability to Innovate: Can the organization sustain value creation as conditions change?
These four lenses offer far more clarity than simply checking if a team is busy or a project is "on track."
They also reframe how we view resources:
Capital is finite.
Talent is limited.
Time is short.
Leadership attention is always stretched.
Every investment is implicitly a decision about what you are not investing in. Product management isn't just deciding what to build—it’s deciding what not to build.
Many leaders treat Agile as a simple choice between operational frameworks: Scrum, Kanban, Waterfall, or SAFe?
That debate usually happens too early. First, get clear on the nature of the problem:
What do we know for sure versus what are we assuming?
Is the operating environment predictable or fast-changing?
How quickly can we get feedback?
How costly is it to be wrong?
Frameworks like Cynefin and the Stacey Matrix are invaluable here for mapping the complexity of a problem before picking a methodology:
Predictable problem? You don't need a heavy adaptive process.
Truly complex problem? You cannot simply impose a rigid linear project plan and expect success.
Match the way of working to the problem at hand—not the other way around.
Having these discussions with senior project managers, IT directors, and transformation leaders can expose a fundamental divide.
Two people can look at the exact same initiative through completely different lenses:
Lens A: Requirements, milestones, deliverables, timelines.
Lens B: Value, evidence, user feedback, opportunity cost, and pivot decisions.
Shifting from Lens A to Lens B requires changing how leaders actually make decisions. It requires getting comfortable saying:
"We don't know yet—let's test that assumption."
"The evidence doesn't support continuing this initiative."
"Users aren't deriving the value we anticipated."
"This is no longer the best use of our capital."
These are tough calls in any organization—especially when reputations, budgets, and executive sponsorships are on the line.
Whether in public sector, enterprise tech, humanitarian work, or digital transformation, the underlying question remains the same:
"Are we generating sufficient value for what we are investing?"
Value isn't strictly financial ROI. Value can mean:
Enhanced service quality
Risk reduction
Improved customer experience
Organizational resilience
Speed and agility
Expanded market access or compliance
However, you cannot measure value without first defining it.
EBM isn't just a metric-gathering framework for Agile teams—it's a model for better investment conversations across the entire strategic chain:
If the chain breaks, you investigate why. And sometimes, the highest-value decision you can make is to stop and redirect resources elsewhere.
The certification itself was just a milestone back in 2021. The real learning has been putting it into practice every day since—across digital transformations, AI initiatives, ERP modernizations, and new product builds.
It equips leaders to shift their core line of inquiry:
❌ “Can we deliver this?”
❌ “How many features did we ship?”
❌ “Are we following the process?”
❌ “How do we keep the project moving?”
Our job as leaders isn't simply to build more things.
Our job is to create maximum value with the finite resources we’ve been given.
That is the difference between delivery and true leadership.
What is Evidence-Based Management (EBM)?
An approach developed by Scrum.org focused on using goals, evidence, and continuous experimentation to steer organizational investments toward real customer and business value.
What are the 4 Key Value Areas in EBM?
Current Value, Unrealized Value, Time-to-Market, and Ability to Innovate.
Is EBM only for Agile coaches?
No. It is built for executives, product leaders, program managers, and portfolio decision-makers shifting focus from activity to outcomes.
💬 Over to you: How is your organization currently measuring the success of major tech and product initiatives? Are you tracking delivery metrics or value created? Let’s connect in the comments.
Every failed project I've worked on traces back to the same place. Not execution. Not the Gantt chart. Somewhere in the six weeks before anyone opened a project plan, when nobody agreed on what problem they were actually solving.
I've watched this play out enough times to stop being surprised by it.
Here's one that stuck with me. A team I led was processing teacher stipends across 123 local school boards. Error rate: 35 per cent. Not a typo. Over a third of payments were wrong in some way, and everyone downstream had learned to live with it.
The instinct in that situation is to jump to the fix. New system, new form, new checklist. We didn't do that first. We spent time figuring out who was actually affected, where the errors were originating, and why the same mistakes kept repeating across boards that had never talked to each other. Turned out the problem wasn't the process on paper. It was what people were actually doing versus what was documented, and those two things had drifted apart years earlier.
Once that was clear, the fix was almost boring. We got it down to 5 per cent with a team of 10.
I bring this up because it's the whole argument, compressed into one number. Projects don't go bad in delivery. They go bad the moment someone skips the problem definition and reaches for a solution instead.
There's a distinction I see missed constantly: the client and the user are not the same person.
The client funds it and carries the reputational risk. The user is the one who has to live with whatever gets built, every day, long after the client has moved on to the next priority. They rarely see the same problem the same way. If only the client is in the room when you're defining scope, you'll build something that satisfies the person paying for it and fails everyone actually using it.
I've seen this most clearly in disaster response, oddly enough. Working relief operations after Hurricane Matthew, Hurricane Maria, and later during the La Soufrière eruption and Hurricane Beryl, there's no version of this work where you plan from a boardroom. You go to where the need is. You listen to the people receiving aid, not just the people coordinating the response from a distance. The gap between what's assumed and what's actually happening on the ground is where relief efforts either work or quietly fail.
That instinct, get close to where the real information lives, isn't specific to humanitarian work. It applies to a stipend process, a policy rollout, a piece of software. Anywhere the person deciding is not the person living with the decision.
In fast-moving field environments, this isn't optional; it's survival. Conditions change hourly. A plan built on Monday's information is often wrong by Wednesday, and the teams that struggle are the ones treating the original plan as fixed instead of as a working draft.
You need enough planning upfront to move. You don't need all of it.
The healthier version of this is constant, small-scale testing against reality. Get something real in front of real people fast, then adjust. Software calls this a minimum viable product. It's not a software-only idea. It's just the discipline of not waiting until everything is finished to find out you got something wrong.
Changing course midway isn't failure. It's what happens when you get better information than you started with. I'd argue treating a course correction as embarrassing is a bigger risk to a project than the correction itself.
The part people skip: inspecting the team, not just the deliverable.
This one took me longer to fully appreciate. It's not enough to inspect the product. You have to inspect how the team is actually working.
That requires real psychological safety, not the version that gets mentioned in a values statement and never practised. People need to be able to say "this is slowing us down" or "I'm blocked" without it costing them something. If that space doesn't exist, problems get hidden until they're expensive, and by the time leadership hears about them, the fix costs ten times what it would have earlier.
A single retrospective scheduled for the end of an 18-month effort is nearly useless. The lessons have already been paid for in full by then. Retrospectives need to happen along the way, while the fix is still cheap.
Define and get agreement on the problem or opportunity first, before the solution. Then go beyond just defining the solution, but get agreement from the client, users and the team building the solution on what success looks like. Know the difference between who's paying and who's living with it, and get both in the room. Plan in small, honest increments instead of one large upfront guess. Document only enough to preserve what was agreed and why. And watch the team as closely as you watch the product, because a broken process quietly produces a broken outcome no matter how good the plan looked on paper.
None of this is exclusive to any one industry. I've watched the same failure pattern in public sector IT, financial services, and disaster relief. Different stakes, same root cause.
Projects don't go bad. They start bad, long before anyone notices.
Where have you seen a project actually start going wrong? Not where it blew up publicly, but the earlier point where the real damage was already done. Drop it in the comments; I'd genuinely like to compare notes.
A framework will not fix poor leadership. A strategy document will not make an organisation effective. And having a project management methodology does not mean that projects will succeed.
I have spent more than 25 years working across different environments, including government, enterprise technology, programme delivery, governance and humanitarian response. One lesson has stayed with me: the quality of decisions often determines the quality of results.
I have seen organisations with well-developed strategies, established reporting structures and experienced people struggle to move initiatives forward. I have also seen how easily teams become occupied with meetings, reporting and competing priorities while the work that matters receives insufficient attention.
The problem is not always a lack of effort or technical competence. Sometimes, it is how people approach problems, make decisions and organise the work.
This is where mental models and frameworks become useful.
Thinking before doing
I distinguish between mental models and frameworks, although they work together.
Mental models influence how we understand a situation. They shape the questions we ask, the assumptions we accept and the decisions we make.
Frameworks give us a structured way to examine problems and organise our response.
Consider systems thinking. A decision made in one department can affect budgets, resources, processes and service delivery elsewhere. Looking at the immediate problem without understanding those connections can create another problem somewhere else.
The Theory of Constraints offers another useful perspective. When an organisation struggles to deliver, adding more people, projects or reporting requirements may not solve the problem. The real constraint could be decision-making capacity, inadequate resources, unclear accountability or a dependency that nobody has addressed.
Then there is opportunity-cost thinking. Every organisation has limited time, money and people. Every new commitment competes with something already underway. Leaders must be prepared to decide what should proceed, what should wait and what should stop.
These are practical leadership questions. They apply whether you are sitting in a boardroom, managing a programme or supervising an operational team.
I strongly support the use of established frameworks, but I do not believe leaders should follow them blindly.
I have worked with different delivery approaches, including Agile and traditional project management methods. Each has its place. The mistake is assuming that one methodology will suit every problem.
A predictable implementation may benefit from detailed planning and formal controls. An uncertain initiative may require experimentation, feedback and adjustment before the right solution becomes clear.
This is where the Cynefin framework is useful. It helps leaders distinguish between situations that are straightforward, complicated, complex or chaotic. The response should reflect the nature of the problem. When the problem or scenario is complex, Scrum is the best option. This lightweight approach is founded on empiricism with three pillars: Transparency, Inspection, and Adaptation that enable teams to provide value incrementally, rather than following a rigid, prescriptive methodology.
There is another danger. Organisations can become very good at producing plans, risk registers, dashboards and progress reports without becoming better at delivering results.
A project can be reported as green while important decisions remain outstanding. A committee can meet regularly without resolving the issues that prevent progress. A strategy can contain excellent objectives without a credible plan for delivering them.
Are we managing the work, or are we actually moving towards the result we were supposed to achieve?
The value of challenging our own decisions
One practice I believe more leaders should adopt is red teaming.
Before committing to a major decision, deliberately challenge the proposal. Examine the assumptions. Identify what could go wrong. Ask what evidence might contradict the recommendation.
Steelman the opposing argument. That means understanding and presenting the strongest reasonable case against your own position before deciding whether it stands up to scrutiny.
This requires intellectual honesty.
Leaders sometimes become too attached to their proposals. The more time, money or political capital invested in an initiative, the harder it can become to acknowledge that the original assumptions were wrong.
But a good decision is not one that survives because nobody challenges it. It is one that can withstand serious examination.
I would rather have a team challenge a proposal before approval than discover six months later that nobody was prepared to question it.
I believe there is a common leadership responsibility at every level: understand the objective, make sound decisions and take responsibility for results.
The application differs.
At board level, leaders need to understand strategy, governance, risk appetite, investment decisions and organisational performance. Their responsibility is to provide direction and appropriate oversight.
Executives must translate that direction into priorities, resource decisions and organisational plans.
Programme and project leaders must turn those priorities into deliverable work, manage dependencies and identify problems before they become serious.
Operational managers and supervisors must understand what needs to happen, who is responsible and when intervention is necessary.
These responsibilities cannot operate independently. If the board approves priorities that the organisation cannot deliver, something is wrong. If executives cannot translate strategy into manageable work, something is wrong. If operational teams are busy but cannot explain how their work contributes to organisational objectives, something is wrong.
Good governance connects these levels. It also establishes who has the authority to make decisions, who is accountable and when an issue needs to be escalated.
I would encourage leaders to develop six areas of thinking:
Strategic thinking: Understand what matters, why it matters and what deserves priority.
Systems thinking: Recognise how decisions in one area affect the rest of the organisation.
Critical thinking: Question assumptions, examine evidence and consider alternatives.
Decision thinking: Understand risk, uncertainty, trade-offs and consequences.
Execution thinking: Translate decisions into assigned responsibilities and measurable results.
Adaptive thinking: Recognise when an approach is not working and change it.
Nobody needs to master every management framework. Nor should organisations introduce a new methodology every time a problem arises.
What matters is knowing which approach fits the situation, understanding its limitations and having the judgement to change course when necessary.
Mental models help us think. Frameworks help us organise our thinking. Governance establishes accountability. Management disciplines help us get the work done. Performance evidence tells us whether we are achieving the intended results.
None of these replaces sound leadership judgement.
How many decisions in your organisation are made because they are supported by evidence, and how many are made because that is how things have always been done?
How many initiatives are genuinely priorities?
How many are competing for the same people and resources? And ;
How often do leaders stop to question whether the original approach is still appropriate?
These are questions worth asking at every level of an organisation.
I believe leadership development should place greater emphasis on how people think, not simply on the number of qualifications, methodologies or frameworks they know.
The real test of leadership is not how many frameworks you can name. It is whether 0you can understand the problem, make a sound decision, get people moving in the right direction and take responsibility for the result.
I explore these ideas through my work in strategic leadership, governance, programme delivery and organisational performance.
If these are the issues you are working through in your organisation, you can find out more about my work and approach at brenthector.com.
Navigating a waterfall-dominated environment as an Agile practitioner often feels isolating. The structured, phase-gated approach of Waterfall features defined stages of initiation, planning, execution, monitoring, and closure. While methodical, it can also be quite rigid. In contrast, Agile thrives on adaptability, collaboration, and iterative delivery. This inherent philosophical clash creates tension: Waterfall demands meticulous planning and strict timelines, while Agile emphasizes flexibility, continuous feedback, and rapid iteration.
Waterfall methodologies rely on linear progression, extensive documentation, and adherence to predefined schedules. Agile, however, prioritizes adaptability, value-driven delivery, and team empowerment. For an Agilist in a Waterfall setting, this dichotomy can lead to friction, where Agile principles may appear disruptive to entrenched practices.
Resistance to Change: Teams accustomed to rigid workflows often view Agile practices as chaotic.
Limited Autonomy: Centralized decision-making stifles the self-organizing teams central to Agile.
Fixed Mindset: Stakeholders often resist changes to the plan, conflicting with Agile’s adaptive approach.
Misaligned Metrics: Success metrics rooted in timelines and budgets overlook Agile’s focus on value and responsiveness.
Finding Common Ground
To integrate Agile within a Waterfall framework, focus on aligning Agile practices with existing structures.
For example:
Combine Waterfall’s emphasis on clear requirements with Agile’s user stories.
Embed iterative reviews into milestone check-ins.
Position Agile as a complement to enhance, rather than disrupt, workflows.
Akbank: By adopting Scrum, Akbank enhanced customer satisfaction and responsiveness, demonstrating Agile’s value in traditional sectors.
Boa Vista: Transitioning to Agile helped this Brazilian credit bureau shorten release cycles and improve team coordination.
Intralinks: Agile adoption improved responsiveness and product delivery, fostering innovation.
Penta Technologies: Initially resistant, Penta’s embrace of Scrum led to a cultural shift toward collaboration and customer focus.
Cisco: Implementing SAFe reduced defects by 40%, improved morale, and increased efficiency in their Subscription Billing Platform.
Delivery Speed: Reduced cycle times signal successful Agile
Adoption.Quality: Lower defect rates reflect improved processes.
Collaboration: Surveys and feedback can measure enhanced team communication.
Customer Satisfaction: Improved feedback scores indicate value delivery.
Employee Engagement: Positive morale shifts suggest cultural alignment.
Regular retrospectives provide ongoing opportunities to evaluate and refine integration efforts.
Leadership Buy-In: Secure commitment from leadership to champion Agile practices.
Incremental Changes: Start small with pilot projects to demonstrate quick wins.
Training and Coaching: Equip teams with the knowledge and support needed for a smooth transition.
Communicate Benefits: Highlight how Agile addresses specific pain points in the Waterfall process.
I keep coming back to a problem that affects me personally, but I think it is much bigger than any one organization.
We continue to put technically capable people into leadership positions and assume that because someone is intelligent, highly skilled or successful in their profession, they will know how to lead people.
That assumption is wrong.
A person can be an excellent engineer, developer, analyst, lawyer or specialist and still be a poor manager.
Technical competence and leadership competence are different things.
The problem becomes serious when someone with poor emotional intelligence is given authority over other people. They cannot handle criticism. They see disagreement as a personal challenge. They rely on intimidation or bullying. They struggle with different personalities. They do not create an environment where people feel safe enough to question decisions, admit mistakes or offer a different view.
The damage does not stop with employee morale.
People stop contributing ideas. Creativity declines. Collaboration suffers. Good people leave.
Some leave physically. Others leave emotionally first.
They remain on the payroll, but they stop caring about the outcome. They stop offering ideas. They do what is required and nothing more. Meanwhile, they quietly look for another opportunity.
Most organizations do not have a real-time measure for that loss.
By the time someone resigns, the organization may have been losing their knowledge, creativity and commitment for months.
That raises a basic question about how we promote people.
The Peter Principle describes the idea that people in hierarchical organizations can be promoted because they succeed in one role until they reach a role where they are no longer competent. I think this becomes particularly relevant when technical excellence is treated as evidence of leadership capability.
Being good at the work does not automatically mean being good at leading the people who do the work.
Management requires a different set of capabilities. Listening. Self-awareness. Communication. Judgment. Conflict management. Empathy. Humility. The ability to develop other people rather than compete with them.
These are often dismissed as "soft skills."
I think that is a mistake.
High Emotional and Social Intelligence are leadership capabilities.
This does not mean IQ or technical competence do not matter. They do. A leader needs enough technical understanding to make sound decisions and ask intelligent questions.
But intelligence does not compensate for poor emotional maturity.
In some situations, it can make the problem worse.
An intelligent person with authority who cannot admit that they are wrong can rationalize almost anything. They can construct sophisticated arguments to defend a bad decision. They can silence disagreement while convincing themselves that they are simply maintaining standards.
That is dangerous.
It becomes even more dangerous as we move deeper into digital transformation and AI adoption.
We are building systems that can process information, generate content, automate decisions and increase the speed at which organizations operate.
But speed does not make a bad decision a good decision.
AI can help us execute a good assumption faster. It can also help us execute a bad assumption faster.
That is one of my concerns about the future.
We may not simply produce faster. We may reach the consequences of our mistakes faster.
Complex problems require people who can question assumptions, test ideas, bring different perspectives and change direction when the evidence says they should.
Psychological safety does not mean removing accountability. It does not mean everyone gets their way. It means people can raise concerns, challenge an assumption, ask a difficult question or admit a mistake without expecting humiliation or punishment for speaking up.
A leader who creates that environment does not weaken authority.
They improve the quality of information reaching the person with authority.
That distinction matters.
If people are afraid to tell the boss that something is wrong, the boss does not have control. The boss has incomplete information.
The scientific method provides a useful example. Science works through questions, hypotheses, experimentation, evidence and revision. A hypothesis can be wrong. When evidence contradicts it, we change the hypothesis.
Leadership requires some of the same intellectual humility.
A leader must be able to say:
"I was wrong."
"The evidence has changed."
"This approach is not working."
"Let's try something else."
That is not weakness. It is part of responsible decision-making.
It requires personal work.
It means listening more.
Being teachable.
Recognising your own biases and blind spots.
Being willing to unlearn.
Understanding how your experiences, fears and assumptions affect the way you treat other people.
For people of faith, the example of Jesus also provides a powerful model of leadership based on service, humility and developing others. Leadership is not simply the exercise of authority over people. It is the responsibility to help people become capable of doing what they could not accomplish alone.
That principle applies well beyond religion.
You cannot hire intelligent people and then create an environment where they are afraid to disagree with you.
You cannot ask people to innovate while punishing every failed experiment.
You cannot say that people are your greatest resource while tolerating managers who consistently drive talented people away.
And you cannot build better technology while ignoring the human behaviour surrounding that technology.
This is why I believe emotional intelligence needs to be treated as a core leadership capability.
Not because technical skills are unimportant.
Because leadership is ultimately exercised through relationships between human beings.
We need leaders who can listen.
Who can challenge themselves.
Who can admit when they are wrong.
Who can create psychological safety while maintaining accountability.
Who can bring different kinds of people into the conversation and actually listen to what they say.
Who understand that authority can make people comply, but it cannot make them care, think creatively or tell you when you are wrong.
The question we should ask when someone is being considered for leadership is therefore bigger than:
"Are they smart enough?"
We should also ask:
"Are they emotionally mature enough to lead other human beings?"
Because the future we are building with AI will still be built by people, for people.
And the quality of those people, especially the people we put in positions of authority, will shape what that future looks like.
Day 1 of the 1st CTU Caribbean Artificial Intelligence Forum, and the panel that stuck with me was "National AI Movements: Trends, Opportunities & Lessons Learnt."
The message was clear.
The Caribbean's AI transformation won't be decided by technology. It'll be decided by whether governments can build the right environment for innovation, governance, investment, and delivery.
Kirk Sookram, Deputy CEO of TATT, moderated the panel. He brought together policymakers, practitioners, and researchers to talk through where national AI initiatives actually stand across the region, what's working, what isn't, and where regional collaboration through knowledge sharing, capacity building, and policy alignment can move things faster.
One point stood out to me.
A lot of governments are trying to run digital transformation and AI programmes through procurement frameworks and delivery models built for a different era. The old playbook assumes you can define requirements fully upfront, pick a solution through detailed specs, and call it success if you deliver against the original plan. That works fine when the problem is well understood and the solution is predictable.
It comes with uncertainty, constant learning, shifting user needs, moving technology, and new risks around data, security, ethics, and governance. The real challenge isn't rolling out a platform. It's changing how public institutions operate and deliver value to citizens.
That's the case for Agile in national digital and AI transformation.
Agile doesn't strip out governance or accountability. It builds a delivery model that lets governments move with discipline while staying open to new information, testing ideas, learning fast, engaging users, and delivering value in pieces instead of waiting years for a final product that might miss the mark by the time it lands.
None of this means Caribbean governments should ditch procurement or programme management. Those disciplines still protect public resources and keep things transparent. What needs to change is how we apply them.
Governments are already starting to move this way. Modular procurement, outcome-based contracts, phased funding, iterative delivery, all working toward the same goal: let the programme learn, adjust, and prove value before you commit more money.
The evidence backs this up. The OECD points to user-centred, iterative approaches as a cornerstone of good digital government. NIST's AI Risk Management Framework makes the same case from a different angle: AI systems need continuous lifecycle governance, not a one-time rollout. Even PMI's own body of knowledge has shifted this way. Three different institutions, one conclusion. Rigid, linear delivery models can't manage digital transformation on their own anymore.
Start with uncertainty. Traditional procurement assumes you can spec out requirements in detail before work starts. National digital transformation almost never works that way. Services shift based on user research, policy changes, new technology, cyber risk, and citizen feedback. AI makes it worse, because model performance, data quality, governance, and regulation can all move mid-implementation. Fixed-scope procurement locks governments into a solution before anyone fully understands the problem.
We can only discover our way out of uncertainty; we cannot plan our way out!
Then there's the track record. Large waterfall ICT programmes have a rough history: cost overruns, slipped timelines, outright failures. The Standish Group's research keeps landing on the same pattern across editions, even as the exact numbers shift. Bigger, more complex projects succeed less often than smaller, incremental ones. Independent reviews out of the UK, Australia, Canada, and the US tell a similar story.
The Standish Group research has consistently shown that large complicated IT projects have considerably lower rates of success than smaller incremental initiatives. While specific percentages differ by edition, project size and complexity remain significant risk factors. Independent reviews in the UK, Australia, Canada and the United States have come to this conclusion.
AI adds another wrinkle. Most traditional projects end at delivery. AI systems don't. They need ongoing monitoring, retraining, bias testing, cybersecurity, regulatory compliance, performance measurement, and human oversight for as long as they're running. NIST is blunt about it: AI needs lifecycle governance, not a one-time implementation. Treat an AI programme like a finite project and you've already misread what you're managing. It's an operational capability.
Procurement rules don't help.Public procurement leans hard on fixed specs, lowest evaluated cost, heavy compliance paperwork, and long approval chains. That protects public money, but it also kills flexibility right when you need it most. The OECD's answer is innovation procurement: outcome-based contracts, modular structures, iterative delivery, used where it makes sense.
And underneath all of it, the real challenge usually isn't the technology. It's the organisation. Political leadership, legislation, funding that survives more than one budget cycle, institutional reform, workforce capability, data governance, getting stakeholders to actually adopt the thing. McKinsey and others keep finding the same root cause behind failed transformations: organisational and cultural factors, not the tech stack.
That's the shift already underway. Procurement that rewards outcomes over fixed specs. Programme management that makes room for experimentation and continuous improvement. Leaders who treat digital transformation as organisational change first and technology second.
The Caribbean doesn't have to repeat mistakes other regions already made. We can take the lessons from global digital transformation efforts and adapt them to our own scale, resources, and regional ties.
AI is coming for government and society either way. The real question is whether our institutions are ready to govern it, adopt it, and deliver it responsibly.
I spend my time helping organisations close the gap between strategy and execution, combining digital transformation expertise, Agile delivery, governance frameworks, and hands-on programme management. The point is simple. Turn ambitious digital and AI strategies into outcomes people can actually see, in services that work better and value that lasts.
The Caribbean doesn't need to wait for AI's future to arrive. We need the capability to shape it ourselves.
#DigitialTransformation #AI #AIGovernance #Agile
In the book "Mastering Professional Scrum: A Practitioner's Guide to Overcoming Challenges and Maximizing the Benefits of Agility (The Professional Scrum Series), by Stephanie Ockerman and Simon Reindl. the writers assert that, apart from being self-organizing, cross-functional, and collaborative, a high-performing Scrum team is stable. Furthermore, as delineated in the Scrum Guided 2020 it is the responsibility of the Scrum Master to act as a servant leader to guide the team (and in some instances, the organization) through the journey of Forming Storming, Norming, and Performing, to accomplish the goal.
Furthermore, to make empiricism work, Scrum Teams need to collaborate to deliver valuable solutions to complex problems, then measure the results, and subsequently adapt based on feedback.
Thus, a servant leader project manager focuses on facilitating cooperation and collaboration, rather than on managing a group of individuals.
Facilitators inspire everyone to achieve their best thinking and work output when working together. Facilitators increase team involvement, comprehension, and shared accountability for team output. Facilitators help the team come up with acceptable solutions.
Servant leaders encourage cooperation and communication both within and among teams. As such, a classic role of the servant leader is assisting in identifying and communicating bottlenecks inside and across teams.
In Summary, to successfully transition to an agile project manager, or to be successful as a project manager in an agile environment, the aspirant should start with a focus on "Individuals and interactions over processes and tools".
Brent is an aspiring Agilist and lover of coffee!
https://www.linkedin.com/in/brent-c-hector/
Chief Engineer Personal Log. Stardate unlisted. Planet: Public Service.)
Every Captain eventually demands more speed.
Scotty and La Forge have never once said yes without a complaint first.
Turns out neither does anyone running a public sector programme.
Here’s the lore translated, dialed up to maximum hilarity.
Scotty (TOS): "The engines cannae take it, Captain! She's gonna blow!" (Oh FYI if you are not going to read this as Scotty, then what's the point?)
La Forge (TNG): "Captain, we are already at 120% of recommended tolerance. The plasma conduits are beginning to fracture."
Public Sector Translation:
"Captain, the caseworkers cannae take it. If we push this rollout any faster, the entire service delivery matrix will experience a catastrophic morale breach. The compliance containment field is already flickering."
What this actually means: frontline staff, IT systems, and approval workflows all have hard throughput limits. Ignore them and you don't get warp speed. You get a workforce core breach. Otherwise known as mass resignations and an Auditor General report.
Scotty: "I cannae do it! I don't have the power!"
La Forge: "We could try, but we'd have to drain power from shields and life support."
Public Sector Translation:
"We can hit your deadline, but only if we divert resources from non-essential systems. By which I mean we're shutting down the policy shop, the stakeholder engagement holodeck, and the monthly 'synergy workshop.' Also, no professional development this quarter."
The workaround: they reroute. Scotty pulls power from phasers. La Forge reconfigures the EPS grid. In bureaucracy, this means temporarily cannibalising other programmes and telling their directors, "Sorry, your project just became a battery."
Scotty: "The crystals cannae take another jump!"
La Forge: "The dilithium matrix is destabilizing. We need to recrystallize."
Public Sector Translation:
"The one person who actually knows how the grant programme works is on stress leave. The legacy database running the payments is written in a dead language from 1987. We cannot go to warp until we fix the crystal, or find a new one."
What they do: Scotty crawls into a Jefferies tube with a wrench. La Forge builds a makeshift recrystallization chamber in a cargo bay. In government, this is the heroic individual who knows the workaround, the jury-rigged spreadsheet macro, or the retired civil servant you un-retire on contract to stop the whole thing from seizing up.
Scotty tells Kirk a repair will take two weeks. Does it in two hours.
La Forge is more honest, but still pads estimates because he knows Riker will ask for it faster anyway.
Public Sector Translation:
"The procurement cycle alone is 16 weeks. The Treasury Board submission is a 9 month orbital decay. You need it by fiscal year end? Fine. I'll invert the tachyon field."
The workaround: Scotty keeps a hidden buffer in his estimates so he looks like a miracle worker the moment the Captain starts shouting. In the public sector, this is called managing upwards. La Forge's version is more modern. He builds an automated script that does the compliance checks in hours instead of weeks. Both are survival mechanisms.
Scotty reluctantly agrees to disable the safety valves.
La Forge: "If I drop the magnetic constrictors, we might get more speed. But if the containment field collapses..."
Public Sector Translation:
"Yes, Captain, I could bypass the privacy impact assessment and the accessibility review. But if we hit an anomaly, say a lawsuit or an access to information request, we will experience an immediate and total ethical warp core breach. The department will be venting plasma into space."
What they do: they don't blindly disable safety, they monitor the hell out of it. Scotty keeps one eye on the pressure gauge. La Forge has Data running continuous diagnostics. In programme delivery, this is parallel track risk monitoring. You speed up the rollout, but your best analyst is watching the lawsuit asteroid field the whole way through.
When all else fails, both engineers have one last move. Creative interpretation of the technical manual.
Scotty rewrites the laws of physics with a hammer and optimism. La Forge inverts a tachyon pulse through the main deflector dish to solve a problem the manual says is impossible.
Public Sector Translation:
"We cannot change the legislation, Captain. But we can issue a policy interpretation that creates a temporary subspace corridor through the regulatory framework. Technically legal. Definitely not what the drafters intended. It will hold together just long enough to reach the next fiscal quarter."
In Star Trek and in the public sector, the engineer's job was never to say no.
It's to say: here is exactly what will break, here is what I need to keep it from breaking, and here is the incredibly risky workaround if you ignore me anyway.
The difference? Scotty does it with a flask of Scotch and a twinkle in his eye. La Forge does it with a VISOR full of diagnostic data and a worryingly detailed contingency plan.
Either way, when the political bridge crew demands warp 9 to meet an arbitrary election cycle deadline, the programme delivery engineer is the one in the engine room, holding the whole thing together with duct tape, regulations, and sheer bloody minded competence.
"Programme delivery in the public sector, Chief Engineer . The bureaucratic equivalent of keeping a Constitution class starship in one piece. She'll hold. But only because I'm a miracle worker."
(Chief Engineer Personal's Personal Log. Computer, archive this one.)
Which one are you on your current programme? Scotty, buying time with a hidden buffer? Or La Forge, running diagnostics while the deadline screams?
The following was originally intended for a talk and panel discussion that I never had the chance to deliver. I recently rediscovered it and decided to turn it into an article. Please feel free to share any feedback.
Caribbean Small Island Developing States (SIDS) face a hard truth. Disaster resilience is not a strategy. It is the core function of national sovereignty. We used to file safety plans away until storms came. That time is over.
The ability to protect life, territory, and critical infrastructure is non-negotiable. If we cannot defend our people from the next shock, do we truly govern ourselves.
Sovereignty means controlling your borders. But what happens when the ocean rises over those borders? What happens when a hurricane wipes out half your capital? The state loses its monopoly on protection.
After Hurricane Maria, a Parliamentary Representative in Dominica told me that he felt like a spectator in his own country. Foreign troops distributed food. Foreign engineers assessed damage. His government stood by while others took the lead. That feeling erodes trust. It tells citizens their own flag offers less safety than a foreign one.
The Economic Commission for Latin America and the Caribbean (ECLAC) reports our region suffers disproportionate economic losses from natural disasters. Sometimes these losses exceed 10 percent of GDP in a single event. That is not just a number. That is a hospital wing that never gets built. That is a school budget slashed to pay for debris removal.
When we rely on external aid for basic survival, we surrender political authority. We become tenants in our own homes.
We face three different resilience problems. Physical damage. Fiscal shock. Digital failure. Each requires a different instrument. We cannot solve a fiscal shock with a concrete wall. We cannot solve digital failure with a catastrophe bond.
We must invest before the disaster. Prevention, mitigation, preparedness and recovery capacity must be funded in advance.
The UN Office for Disaster Risk Reduction states resilient infrastructure is an investment in the continuity of critical services. We cannot wait for the roof to blow off to fix the foundation. We need building codes that hold. We need evacuation plans that work without outside help.
Fiscal autonomy is impossible when we borrow to fix the same broken things every few years. The cycle traps us. A storm hits. We borrow billions. We rebuild. The storm hits again. We borrow more.
Our economic policy becomes dictated by creditors rather than national needs. We need to expand regional risk pools like the Caribbean Catastrophe Risk Insurance Facility (CCRIF). We must use catastrophe bonds and other capital market instruments to transfer extreme sovereign disaster risk.
This shifts disaster financing from improvised panic to planned execution. Money must be available immediately.
Connectivity is territory. A severed undersea cable isolates a nation faster than a naval blockade.
Hurricane Beryl in 2024 exposed the vulnerability of communications infrastructure across the Eastern Caribbean. Severe outages hit Grenada and Saint Vincent and the Grenadines. Banks could not process transactions. Government alerts went unsent. Families could not confirm safety. The nations went mute.
A data centre by itself is just a concrete box with servers. A resilient digital operating environment is a living nervous system. Critical platforms, applications and government data must be replicated across geographically separated environments. We need resilient power, tested backups and defined recovery time objectives. NIST contingency planning guidance explicitly requires these alternate processing locations and regular testing.
This is where regional coordination matters. CANTO Caribbean brings telecommunications providers together. The Caribbean Telecommunications Union (CTU) sets policy frameworks. The Caribbean Datacenter Association (CDA) ensures physical data infrastructure survives climate shocks.
There is only one test that matters. If a Category 5 hurricane hits tomorrow, how quickly can the country restore electricity, connectivity, identity, payments, health and emergency communications.
And where is the money already available to do it.
The CARICOM Strategic Framework for Regional Digital Resilience recognizes this exact priority. A hurricane that destroys a government building is a physical disruption. A disaster that takes down identity systems, payment platforms and health records is a national digital continuity crisis.
I think about the children born this year. What kind of sovereignty will they inherit. Will they govern nations that stand firm. Or will they manage territories that constantly beg for help?
We have the technology. We have the knowledge. We just need the will to pay for it before the sky turns green. Let us make 2026 the year we stop treating resilience as an emergency expense. Let us make it the definition of our statehood.
What does resilience as sovereignty mean for Caribbean SIDS? It means the ability to protect citizens and infrastructure without relying on external aid. When states handle disasters independently, they maintain political authority.
Why separate physical, fiscal, and digital resilience? They solve different failure modes. Physical resilience prevents damage. Fiscal resilience provides immediate liquidity. Digital resilience ensures critical government services stay online or recover rapidly.
How do natural disasters affect Caribbean fiscal autonomy? Repeated borrowing for reconstruction creates debt dependency. Creditors gain influence over national budget decisions. Pre-arranged financing prevents this loss of control.
Why is digital infrastructure considered sovereign territory? Connectivity enables governance during crises. Without it, states cannot coordinate response, send alerts, or maintain economic function. A digital blackout paralyzes the state.
How can Caribbean states break the disaster debt cycle. Invest in mitigation before disasters. Expand regional insurance pools like CCRIF? Use catastrophe bonds to transfer extreme risk. Treat resilience as a core budget function.
What is the difference between a data centre and a resilient digital environment. A data centre is a physical building? A resilient digital environment includes geographically separated infrastructure, redundant power, tested backups, and defined recovery time objectives to guarantee service continuity.
Economic Commission for Latin America and the Caribbean. 2023. Assessment of the economic impact of natural disasters in the Caribbean 2020 to 2023. United Nations.
World Bank. 2025. Climate Risk Country Profile: Caribbean Small Island Developing States. World Bank Group.
CARICOM. 2025. Strategic Framework for Regional Digital Resilience 2025 to 2030. CARICOM Secretariat.
UN Office for Disaster Risk Reduction. 2024. Principles for resilient infrastructure.
NIST. 2024. Contingency Planning Guide for Information Technology Systems.
Walk into almost any boardroom when an operational problem comes up, and you can predict the next sentence before anyone speaks: What software can we buy to fix this?
Missed deadlines? Let's procure an enterprise project suite. Messy reporting across units? Let's commission an executive dashboard. High error rates in routine administration? Time for a full-scale ERP overhaul, or whatever "intelligent" platform is trending this quarter.
I have spent twenty years leading programme governance, public sector modernization, and system implementations across the Caribbean, and I keep watching the same expensive mistake play out. Buying software to cover up an unclear process does not fix the problem. It just automates the mess at a much higher price tag.
When leaders chase shiny platforms before thinking from first principles, they confuse a software license with operational discipline. In my experience, the fixes with the highest return rarely come from heavy enterprise tech. They come from low-tech, practical changes built by stripping a problem down to its actual bottleneck.
Enterprise software, BI dashboards, and AI tools work well when they sit on top of a clear, well-run workflow. Without that foundation, three things usually go wrong:
You lock bad habits into expensive code. If roles, handoffs, and decision rights are fuzzy on paper, configuring an ERP around them just hard-codes the confusion. Every minor change later becomes slow, costly, and political.
Dashboards turn into expensive wall art. A live analytics screen looks great in a steering committee meeting. But if frontline teams don't agree on how data is defined, or nobody is held responsible when an indicator turns red, nothing on the ground changes.
Maintenance drains the team. Complex platforms bring recurring license fees, vendor dependencies, and constant retraining. Eventually, staff spend more energy feeding the system than serving the public or the customer.
Thinking from first principles means ignoring vendor pitches and asking three blunt questions:
What is the actual bottleneck stopping us from delivering on time?
What is the simplest way to remove that bottleneck?
If we had $0 for new software this year, how would we solve 80% of this problem next Monday morning?
In government agencies, regulated companies, and field operations, the answers are usually surprisingly plain:
Instead of an enterprise workflow suite to fix dropped handoffs: Use one shared tracking board—a structured spreadsheet or even a physical Kanban wall—with named owners, a clear definition of "done," and limits on how much work can sit in progress at once.
Instead of a custom mobile app for field reporting: Build a standardized, five-question form (using accessible tools like KoboToolbox or simple templates) around clean, unambiguous fields.
Instead of a multi-source BI warehouse for executive visibility: Require a one-page Friday status brief tracking three core numbers, backed by a disciplined 20-minute review meeting.
Instead of automated auditing tools to catch recurring errors: Enforce a five-point pre-flight checklist before any file, payment, or deliverable moves to the next desk.
None of these fixes require a seven-figure budget or an 18-month rollout. Yet when paired with clear ownership, they routinely clear out most operational friction in a matter of weeks.
Low-tech does not mean anti-technology. It means getting the order right.
Before you sign off on the next ERP module or dashboard build, ask for proof that the workflow runs cleanly by hand at small scale. Standardize the inputs. Clarify who approves what. Build the habit of follow-through.
Once a process is predictable and owned by the people running it, software becomes a real asset instead of an expensive crutch. Until then, keeping things simple isn't a compromise. It's the smartest move on the table.
Where have you seen a simple checklist, template, or weekly routine outperform an expensive tech rollout? I'd like to hear what worked in the comments.
I’ve spent years right in the middle of it all—watching, leading, and digging into digital transformation as it unfolds, or honestly, stalls in both government and private companies. One thing keeps happening: some big, polished strategy from ITU, ECLAC, UNDP, or UNESCO drops with plenty of hype, then nobody really uses it. The real challenge isn’t about grabbing the latest software. It’s wrestling those messy human and organizational problems that quietly choke progress. Most enterprise AI projects simply don’t get off the ground. You see numbers like 70 to 95 percent failure rates. I’ve looked at why, and every time, it boils down to three main blockers: fear of risk, poor financial reasoning, and cognitive bias.
Fear. Bad ROI logic. Cognitive bias. The three things killing your Digital Transfomation of AI project.
Let’s talk about fear. Not the fuzzy, hard-to-pin-down kind—this is real, practical fear. Leaders look at what could go wrong and decide it’s just not worth the headache. So projects stall out. And they’re not wrong. AI projects fail about twice as often as typical IT efforts. It’s rarely the algorithm’s fault, either. The usual culprits? Messy data, fragile systems, and ignoring change management altogether. Then there’s all the technical debt nobody talks about: models that stop working (model drift), software spaghetti nobody can untangle, and systems that are impossible to support long term. Most teams just aren’t built for model reliability, so those cracks get bigger and bigger.
Government work? Even tougher. Pile political risk on top of all those technical headaches. It’s not just about being efficient—it’s about slogging through bureaucracy, dodging silos, and dealing with leaders who refuse to kill failed projects because admitting a mistake kills their own career. I call this “risk avoidance theatre.” So you get endless meetings, mountains of checklists, and layers of oversight that look like smart leadership but actually grind everything to a halt.
Then there’s the legal and reputational risk. With new regulations, standards and frameworks like the EU AI Act, the GDPR (General Data Protection Regulation), NIST AI Risk Management Framework (AI RMF) or ISO/IEC 42001 (AI Management Systems); a misstep can mean huge fines or damage to public trust. Sometimes, just the whiff of risk is enough to shut a project down.
So how do you move past the fear? I recommend a Pre-Mortem Analysis. Don’t just ask, “What could go wrong?”—tell your team to imagine it’s six months from now, the project has crashed and burned, and now it’s time to figure out why. Pull in the tech, business, legal, and finance folks, give everyone 15 minutes to think it through, then build your risk plan based on what comes up. It encourages honesty and cuts through useless optimism, but you have to facilitate it well or people just write down minor annoyances and miss the real threats. And yes, new risks can still come up later, so stay on your toes.
Risk avoidance theatre: meetings, checklists and oversight that look like leadership but produce paralysis instead.
Then there’s the finance piece. Bad ROI metrics stop projects in their tracks just as much as fear does. Most organizations are lousy at measuring the true value of AI. If the CFO (Cabinet, Ministry of Finance) can’t see clear returns, budgets get frozen, projects stall out, and suddenly their skepticism looks justified. The problem? Teams measure the wrong things—API calls or models trained say nothing about real impact. There’s also the causality trap: it’s tough to tell if your AI tool, or any digital transformation implementation, boosted business results or if some other initiative moved the needle. Plus, companies gloss over the hidden costs—data prep, integration, and fixing bad AI output eat up as much as 40 percent of the gains you expected. And AI’s value compounds over years, while finance wants quick wins every quarter.
To get this under control, set up a cross-functional Value Realization Board and run empirical funding loops. Break big, slow projects into fast, focused experiments. Before you start, anchor every AI project to a concrete business outcome tied to a real line item—like reducing churn by 5 percent, not just “improving customer experience.” Fund projects in 90-day sprints. If teams hit their target with real evidence, fund them again. If not, pull the plug. This stops teams from gaming vanity metrics and gives finance a reason to trust the process. Don’t be surprised if you get some pushback, though—CFOs hate to change how the budgeting sausage gets made. Plus, there’s always a risk that teams chase easy wins or start gaming the system.
You can't sell a a clean ROI story while they're still paralysed by risk and pricing your project against last year's software licence.
Finally, even if you’ve handled risk and ROI, cognitive biases can still quietly wreck your plans. Leaders aren’t robots—they anchor to the first price they see, or the last failed project. CFOs might use old software license costs to judge a new AI system, completely missing all the new ways it delivers value. Loss aversion is even worse—nobody wants to risk a loss, so they kill off anything that feels uncertain. Status quo bias keeps people clinging to what they’ve got, and overconfidence boosts the chance that both developers and end users trust AI outputs more than they should.
You can push back on these biases with small behavioral nudges. In procurement, ask vendors for multiple pricing models to stop people from anchoring on just one number. When fighting loss aversion, show your staff what they’re missing out on—“People using this tool save five hours a week; you could too.” For status quo bias, make the old way a little more annoying so the new AI system feels easier. Nudges are cheap and mostly respectful of autonomy, but the effect fades if people feel like you’re trying to manipulate them.
Make the old process a little more annoying so the new one feels like the easy choice.
The bottom line: getting from AI vision to real business value is hard, with obstacles on every side—risk, ROI, bias. Fixing just one won’t cut it. You can’t sell a bulletproof ROI story to a CFO who’s paralyzed by risk and still thinking about last year’s costs. My job as a delivery leader is to keep pushing through all this. I lean on pre-mortems, empirical funding, and behavioral nudges to keep things moving. Those big international blueprints? They’re just the beginning. In your first 100 days, focus on building the right structures, not writing endless reports. Stand up a central delivery unit, grab your first round of funding, and launch a small lighthouse project to get a visible win. Share honest results—especially when things aren’t perfect. That’s how you build real trust and how you actually get things done.
I think I'm approaching that stage of life where I find myself reflecting more often. Not quite the old man sitting in a café telling stories to anyone who will listen, but old enough to start recognizing patterns that weren't obvious when I was younger.
One of those patterns came from an unexpected place: project management.
Most project managers learn early about the triple constraint — time, cost, and scope. Change one, and the others move. Want more features? It costs more or takes longer. Need it faster? Cut scope or throw more resources at it. The whole challenge is balancing those competing pressures without wrecking quality.
But life doesn't work that way.
In life, there is really only one constraint that cannot be negotiated: time.
Money can be earned, lost, and earned again. Opportunities close and others open. Skills can be developed. Even some relationships can be repaired. But time goes in one direction. We each get twenty-four hours a day, and none of us knows how many days we've been given.
Once I accepted that, my thinking started to shift.
How I spend my time became a more pressing question than how I spend my money. Commitments started to carry more weight. Every yes became an implicit no to something else. I realized that overcommitting wasn't ambition — more often, it was just a failure to take scarcity seriously.
King Solomon said it plainly in Ecclesiastes: there is a time for everything. A time to plant and a time to harvest. A time to weep and a time to laugh. A time for war and a time for peace. The wisdom isn't only that these seasons exist. It's knowing which season you're in and having the sense to act accordingly.
Right now, Brazil is playing Haiti. I am in the back room playing hide and seek with my four-year-old granddaughter.
Not a distraction. A choice.
When time is the primary constraint, priorities stop being abstract.
You become more deliberate about who you invest in. You look for work that actually creates value. You stop chasing activity and start looking for meaning. You learn that a boundary isn't selfishness — it's the thing that makes your yes worth anything.
For me, the goal isn't happiness. Happiness moves around. What I'm after is something steadier — a life of meaning, service, and real joy. To love God is to love people. To serve others well requires being a responsible steward of the time you've been given.
Maybe wisdom is just that: understanding that time is the one thing you cannot get more of, and choosing accordingly.
There is a familiar ceremony in government offices.
A minister stands at a podium with a new software system projected on the screen behind them. Officials, donor representatives, and visiting consultants gather in the room. Someone cuts a ribbon, or clicks a mouse for the cameras. Everyone applauds. A press release announces a groundbreaking new chapter for service delivery.
On Monday morning, civil servants return to their old routines.
In There's No Such Thing as an IT Project, Bob Lewis and Dave Kaiser describe a pattern common across public institutions. When organizations frame operational reforms as IT projects, they rarely achieve actual change. Software gets delivered, but operational habits stay the same.
Reading their work puts words to a reality many public servants observe daily.
Lewis and Kaiser reference the classic IT cartoon of a tire swing modified by different departments until it becomes unusable. The final panel usually shows what the user wanted: a simple tire on a rope.
The authors reframe this image. The final panel still misses the target because the desired outcome was never the swing. The outcome was children playing on it.
In Caribbean public institutions, agencies frequently fund systems that go unused. Social service agencies install case management software while files remain on paper. Ministries adopt financial platforms while budget officers keep maintaining separate working spreadsheets. Citizen portals launch and log forty transactions in six months because no one tested the workflow with actual citizens.
Projects measure deployment instead of use. They treat launch day as the finish line.
I encourage you to pick up a copyReading it felt like someone had finally written down the thing we all knew but kept not saying out loud.
Lewis and Kaiser outline a structural dynamic they call a "covert conspiracy of failure." Both technology teams and executive leadership meet their immediate goals even when the system yields no operational benefit.
The sponsoring executive claims credit for modernization. The vendor collects payment. The project manager reports a green status dashboard. The IT team delivers according to specifications. If operational habits do not change, leadership labels it a user adoption issue.
This outcome stems from structural incentives. Capital projects often rely on international donor funding with disbursement conditions tied strictly to procurement and installation milestones rather than long-term operational performance. The evaluation criteria focus on whether software was installed and training was conducted, not whether tax processing became faster.
In some cases, key figures prefer incomplete implementation. Full adoption brings direct accountability for operational results, whereas an incomplete roll-out shields leadership from scrutiny.
The book offers a practical shift in approach. Instead of asking "what do you want the system to do?", leaders must ask "how should this department operate differently and better?" The first question treats technology as a vendor; the second treats it as an operational partner.
Consider a security infrastructure project at a regional financial institution. The initial requirements document was copied directly from an overseas parent company. The parent company operated on cloud infrastructure, whereas the regional entity ran on-premises systems. This fundamental difference meant distinct risk profiles, compliance rules, and board approval steps.
Because the team prioritized the requirements document over early operational discussions, procurement moved forward without addressing these differences. Months later, when a project manager finally consulted the local staff who would run the system, they discovered the specifications included features the regional office did not need while omitting essential local requirements.
The failure was not technical; it was a failure to hold substantive operational conversations before selecting technology.
Lewis and Kaiser emphasize that operational failures root in institutional culture. Culture represents the repeated behaviors people adopt in response to their environment. It changes when leadership models new behaviors, not when they issue mandates.
In Caribbean public administration, rigid organizational silos divide government into separate columns, while citizens experience services across rows. Inter-agency collaboration stalls because existing structures do not reward joint work. Risk aversion remains a rational choice when penalties for visible failure are severe. Staff learn to document process steps and label initiatives as pilot programs to avoid risk.
Small island contexts add relational dynamics. In small societies, officials, vendors, and department heads share personal and professional networks. While these connections can build trust, they can also hinder honest feedback when projects misalign with operational goals.
Lewis and Kaiser identify seven core disciplines required for organizational change, including technical architecture, leadership, and project management. The discipline most consistently underfunded in public sector initiatives is organizational change management.
Budgets cover hardware, software licenses, and initial training sessions. They rarely cover the sustained work of guiding staff through operational transitions. When project budgets shrink, change management is typically the first item cut.
As a result, staff return to established habits months after launch because no one managed the operational transition.
Public sector institutions possess significant capability. During emergency responses and natural disasters, government agencies often adapt quickly and coordinate across departments effectively.
Technology projects, however, repeatedly fall into predictable patterns. Institutions fund software delivery rather than operational outcomes and celebrate launch events over functional improvement.
Bob Lewis and Dave Kaiser provide a clear diagnostic framework for these structural patterns.
Before approving project plans or funding allocations, ministers, permanent secretaries, and department heads should ask two questions: How will this specific public service run differently and better? And who holds explicit accountability for achieving that operational result, beyond delivering the software?
Without clear answers to both questions, new systems will continue to produce ceremonies rather than public value.
Brent is a Caribbean-based ICT and governance practitioner with experience spanning public sector reform, disaster response, and organisational change management.
What if your organisation is not struggling to execute because your people are moving too slowly?
What if you are paying them to wait?
Wait for another review.
Wait for another meeting.
Wait for more analysis.
Wait until someone is completely comfortable with a decision that can never be 100% certain.
That delay has a cost.
Lean management recognises over-processing as waste. Kanban focuses on flow and limiting work in progress. The economic question is simple:
What are we getting for the additional time and money we are spending?
This does not mean lowering standards.
If a product must meet a compliance requirement, meet it properly. If additional work reduces a real risk, improves safety or provides something the customer actually needs, spend the resources.
But if the requirement has been satisfied and people continue rewriting the policy because someone wants it to be “perfect”, what is the organisation buying?
The same applies to projects.
Requirements expand.
Features get added.
Documents go through another round of edits.
Decisions are deferred because someone wants more information.
Eventually, the additional effort produces little additional benefit.
That is diminishing returns.
For me, good delivery starts with defining what done means.
What must be delivered?
What quality standard must it meet?
What compliance requirements apply?
What risks must be addressed?
What evidence is sufficient to make the decision?
Once those questions are answered, the focus should be execution.
Not endless refinement.
A leader will rarely have perfect information. The responsibility is to make a proportionate decision using the evidence available, manage the risk, and change course when material new evidence appears.
That is very different from making a reckless decision.
It is also different from using “perfection” to avoid making one.
This is how I approach programme and project delivery.
I help organisations establish the outcome, define what “done” means, identify the work that actually matters, and remove unnecessary delay and expenditure.
My focus is practical:
Define. Decide. Deliver.
If your project is stuck in analysis, requirements keep expanding, decisions are taking too long, or your team is spending money polishing work that is already fit for purpose, let’s talk.
I can help you assess where the delivery process is creating unnecessary cost and where management intervention can get the work moving again.
The objective is not perfect delivery.
It is delivering the required outcome, to the required standard, without wasting resources getting there.
The worst project managers I've encountered had plenty of authority and used it poorly. The best had none and delivered anyway.
There's a myth in programme delivery: your role title determines whether you can get things done. If you're a "Director" you move fast. If you're a "Coordinator" you wait for permissions. It's backwards.
I've coordinated complex delivery across security, compliance, operations, and technology teams where I had no line authority over anyone—not one person reported to me. I had no budget approval rights. I couldn't fire anyone or reassign them. But when a vendor timeline conflicted with a regulatory deadline three months into delivery, I made the call that realigned the entire programme. Not because I had authority. Because everyone in the room understood what the constraint was, what the options were, and why the decision protected the outcome that mattered.
That's not authority. That's credibility born from clarity.
Here's what actually determines whether you can orchestrate complex delivery: Can you translate between worlds that don't naturally speak the same language? Can you take what Compliance cares about, what the vendor cares about, what Operations needs, and articulate why they're not actually in conflict—or if they are, why one has to win and what happens if it doesn't? Can you hold that clarity steady when pressure arrives and people want to renegotiate?
Most project managers think the answer is "get more authority." Escalate higher. Get a bigger title. Build a larger team reporting to you. But authority is brittle. The moment you leave, or priorities shift, or a new executive arrives who doesn't respect your position, it evaporates. Credibility is portable. If you're known as someone who thinks clearly about constraints and makes defensible calls, that travels with you across every role, every organization, every sector.
I've managed competing programmes where projects were fighting for the same resources. The tool that resolved it wasn't a RACI matrix or escalation to senior leadership. It was brutal honesty about what each programme actually needed versus what it wanted, and a willingness to call out when leaders were making priority decisions based on incomplete information. That required no formal authority. It required a reputation for telling the truth about trade-offs.
The practical implication: if you're sitting in a role where you feel powerless because you lack authority, you're solving the wrong problem. Build credibility instead. Become the person who understands the constraints better than anyone else. Become the translator between functions. Become the one willing to say "we can't do both, here's what happens if we try."
Authority gets you a seat at the table. Credibility gets the table to listen.
What's the difference between the most effective coordinators you've worked with and the ones who had titles but got nothing done?
What I really meant at the CTU AI forum
I made a point at day 2 of the 1st CTU AI forum at UWI. I've been turning it over in my head ever since. It needs more honesty than I gave it in the room.
We throw around "shared vision" like it's magic. Everyone nods. Everyone signs off. Six months later we're staring at an AI system nobody uses, wondering what went wrong.
Here's the thing. A vision that is merely shared is just a document that survived a meeting without an argument. It doesn't mean the customs officer who'll use the system was in the room, or that the nurse who'll enter the data had a seat. It doesn't mean the minister who'll defend the budget understood the trade-offs.
We need a vision that's fought over and agreed upon by the people who'll live with the consequences. The room needs to look different. The frontline worker and the decision maker both need a seat. The policy officer and the IT director too. So do the people who'll write the code and the people who'll clean up when it breaks. All of them. Together. Before anyone signs a purchase order.
And that agreement needs to be real. Not a signature on a minutes sheet. A genuine acceptance that this is the direction and these are the trade-offs. And a clear-eyed understanding of what we're willing to risk. Without that, you're not building transformation. You're buying software.
It doesn't stop at the vision. We need a negotiated problem statement too.
Too often the problem statement arrives in a vendor's brochure. It's polished. It's generic. It sounds impressive. And it has nothing to do with the actual friction that citizens feel when they walk into a government office or the bottleneck that keeps an employee at their desk two hours past closing time.
A problem statement that is imposed from above, or imported from abroad, will always misalign with local reality. It will solve a problem that doesn't exist while the real problem keeps festering. You can have the most elegant algorithm on the planet. If it's solving the wrong problem, it's still failure. Just expensive failure.
The problem statement needs to come from the ground. It needs to be recognized by the people who experience the pain every day. When you describe the problem to a frontline officer, they should nod and say yes, that's exactly what happens to me on Tuesday afternoons. If they don't recognize it, you haven't found the problem yet. You've found a hypothesis. And hypotheses need testing, not procurement.
This is where we keep falling down. We ignore people and processes. We rush to install the model, deploy the platform, roll out the digital infrastructure. We get excited about the technology because it's tangible. You can see it. You can demo it. You can take a photo of the launch.
People and processes are messier.
They're harder to photograph and they don't fit neatly into a project timeline. So we skip them. We put the cart so far ahead of the horse it's practically in another parish.
We throw launch parties for systems that haven't changed a single citizen's life. We celebrate going live while the employee is still working around the same broken workflow, except now it's on a screen instead of a paper form. We install AI models that predict things nobody asked to predict. We build digital infrastructure that moves data faster between departments that still don't talk to each other.
If we're serious about bringing AI and digital transformation to the Caribbean, we need to map the citizen journey. We need to map the employee journey. We need to walk through the actual steps a person takes when they need a service, and we need to ask where it hurts. We need to watch how an employee spends their morning, and we need to ask which parts make them want to quit.
Technology should remove friction from those journeys. It should make the citizen's wait shorter and the employee's day smoother. If it doesn't do that, why are we building it? If the answer is "because the grant said we had to" or "because the vendor said it was best practice," then we need to stop and rethink.
Success in our context is too often measured by outputs. We went live. We trained the model. We hit the deployment date. We checked the box.
Those are outputs. They are not outcomes.
An outcome is a citizen who gets their permit in two days instead of two weeks. Or, in my instance, my passport, in one week as opposed to eleven weeks (as of the posting date) and counting. Or a nurse who finally gets to spend more time with patients than paperwork. Or a customs process that catches contraband without treating every honest importer like a suspect.
And here's the thing. Those outcomes need to be negotiated too. What does value actually look like? Who decides? Is it the minister? The permanent secretary? The frontline manager? The citizen? If they don't agree on what success looks like, you'll have one group celebrating while another group is drowning.
Success needs to be specific and visible. And it needs to be agreed upon by the people who'll still be here when the project is over, not by the consultants who'll leave in six months.
Who has the authority to flip the switch and move from pilot to production? Who is writing the policy that lets this thing operate legally? Who is making sure the data protection rules are followed? Who owns the risk when the algorithm makes a bad recommendation? Who is measuring whether the benefits are actually showing up?
These questions aren't decorative. They determine whether your project lives or dies. I've seen too many Caribbean initiatives stall in pilot purgatory. They work technically. But they're politically unsupported and operationally orphaned. Nobody has the clear authority to scale them, and nobody has the clear accountability when they fail.
You need to negotiate this in advance. You need to know who makes the decision. You need to know who has the power. You need to assign responsibility for the regulatory environment, the policy framework, the legislation if that's what it takes. These aren't afterthoughts. They're the foundation. Build the foundation first, or watch the house tilt.
I find that when I give a contribution in a forum or at a meeting, or in a casual discussion, this comment often disturb some people. I warn them that their plans may come to naught, their assumptions may be false, and the terrain may change.
We don't like hearing this. We want certainty. We want a beautiful Gantt chart that stretches neatly from January to December. We want to believe that if we plan hard enough, we can eliminate risk.
You can't. Not when you're dealing with systems where the technology is only one piece of the puzzle. Not in a Caribbean context where institutional capacity changes from year to year, funding cycles shift, and political priorities turn on a dime. You are heading into uncertainty. Accept it.
That negotiated vision you started with? It's a starting line, not a contract carved in stone. That problem statement you fought so hard to agree on? It will need adjusting. The benefits you promised? They'll shift as reality teaches you what's actually possible.
This isn't failure. This is learning. But learning requires feedback loops that actually loop. You need quick feedback. You need to know who gives it. You need mechanisms that get that feedback into the hands of people who can act on it. User councils. Operational reviews. Honest conversations with frontline staff who aren't afraid to say this isn't working.
At the end of all of this, we need to ask a basic question. Are we delivering value to the customer and the citizen? What about the employee behind the screen?
If the answer is no, it doesn't matter how sophisticated your AI is or how many terabytes you've moved to the cloud. It doesn't matter how many dashboards your ministers can see. You have failed.
The focus has to shift to outcomes and value. Get the stakeholders to agree on it. Get the people with budget authority to sign off on it. But never lose sight of the person on the other side of the counter and the employee behind it.
This isn't about installing technology. It's about transformation. It's about changing how organizations work and how citizens experience government. It's about whether employees go home feeling like their day got easier or harder. It's about making sure that when we put something into production, the benefits actually reach the people who were supposed to get them.
We need to stop racing to install and start building things that actually work for people. We need to stop digitizing dysfunction and start designing for humans.
I made my point at UWI. Now I'm elaborating. Let's put the horse back in front of the cart.
Let's build transformation that lasts.
#CAIF #CTU #UWI #CaribbeanAI #DigitalTransformation #Governance #StakeholderAlignment #ProgrammeDelivery #TransformationLeadership
"Agile is a mindset, Scrum is just the how." That line stopped sitting right with me a while back, and here's why.
Scrum is not equal to Agile, fair enough. But calling Scrum "simply the method" gets the history backwards. Before 2001, Jeff Sutherland and Ken Schwaber were already running Scrum. Publicly presented at OOPSLA in 1995, it drew on ideas from Takeuchi and Nonaka's 1986 paper. XP, DSDM, Crystal, all of it predates the Manifesto too. By the time the seventeen authors met at Snowbird, Scrum wasn't waiting to be invented. It was already a decade old. Rather than creating the mindset, the Manifesto gave a name and a shared language to something several frameworks had been practising for years.
Here's my actual thesis, and I'd rather lead with it than bury it: a process tells people what to do. A framework shapes how people think, while leaving room for teams to discover the best way of doing it. Scrum belongs to the second category, not the first.
Look closely at the Scrum Guide. Nowhere does it call Scrum a methodology. Instead, it describes a lightweight framework that helps people generate value through adaptive solutions for complex problems. That's not a process handbook. What must exist gets defined, not every step you must take, and that gap is the whole argument.
Two things sit underneath the framework that aren't procedural at all: empiricism and Lean thinking. Learning from evidence rather than prediction, that's empiricism in a sentence. Make the work visible. Inspect it. Adjust. Optimising the whole system and cutting what doesn't create value, that's Lean thinking doing the same job from a different angle. Neither gets ticked off a list. Both are beliefs about how uncertainty should be handled.
Then come the five values. Commitment, Focus, Openness, Respect, Courage.
Scheduling courage into a fifteen-minute Sprint Planning slot simply doesn't work. Checklisting your way into openness doesn't either. How people behave when the plan falls apart, not whether the sticky notes are colour coded, is where these values actually live.
Mechanical Scrum exists for exactly this reason. Every ceremony runs on time. Every artefact gets filled in. Sprint Reviews happen. So do Retros. Yet somehow nothing changes. Status updates replace genuine Reviews. The same three action items surface in every Retrospective, and none of them get done. Mechanics survive here, philosophy doesn't, and what's left is project administration wearing Scrum's name tag.
Backing this up is Evidence-Based Management, Scrum.org's own guide. Whether you ran the ceremony isn't the question it asks. Where does your greatest unrealised value sit? What assumptions are you standing on? What experiment should you run next? None of those are implementation steps. Ways of seeing a problem, that's what they are.
Rather than Agile producing Scrum, the relationship runs the other way. Existing frameworks like Scrum, XP, Crystal, and DSDM weren't replaced by the Manifesto, they were distilled into a shared vocabulary the whole industry could use. Less a methodology and more a philosophy that emerged from these schools of thought and practices, that's what Agile really is, and Scrum didn't just express it. Scrum helped shape it.
Customer value, learning, adaptability, collaboration, that's the direction Agile points toward. A deliberately constrained operating system for pursuing those outcomes is what Scrum provides. Observable artefacts and frequent inspection impose transparency, it isn't a matter of preference. Required by every Sprint, not offered as a nice-to-have, that's what feedback loops actually are. Teams don't learn on their own time here. Built into the cadence itself, learning simply happens.
That's what occurs when Scrum is reduced to "the how." Agile is a source of philosophy. Scrum is one of the most influential operating systems for adhering to its principle. One is used to describe a destination. The behavior needed to achieve it is shaped by the other.
Where do you draw the line between framework and mindset in your own teams?#CreateImpact