When the Work Changes Hands: AI Workforce Transition Decisions Caribbean Boards Must Own

What wave two does to the people on the payroll — and the workforce decisions a Caribbean board must own before automation moves from pilot to production.

Automation does not arrive as a headcount number. It arrives as a change in who does what — and that change is a board decision, not an IT one.

Every conversation about AI eventually becomes a conversation about jobs. Most Caribbean boards have that conversation too late — and in the wrong room.

This is the fifth paper in Dawgen Global’s Caribbean AI Realisation Series. Edition 01 set the discipline for a first AI pilot. Edition 02 named the five use cases worth funding first. Edition 03 took up wave two — what an enterprise funds second, once a pilot has worked. Edition 04 drew the line between AI that recommends and AI that acts, and set out the authority a board delegates when it lets a system act on the enterprise’s behalf.

This paper takes up the question those four have been circling: what happens to the people. When a pilot becomes a programme, and a programme becomes the way the work is done, the change does not announce itself as a redundancy notice. It arrives as a quiet shift in who does what — which tasks leave a human desk, which arrive at one, and which simply disappear. Handled well, that shift makes Caribbean enterprises more capable and their people more valuable. Handled badly, it produces the two failures we see most often in the region: capability that erodes because the institution forgot how to do the work it automated, and trust that collapses because the workforce learned about the change after the fact.

This paper is about owning that shift deliberately. It introduces a way to map what automation actually does to work — not to headcount, but to tasks. It sets out a four-stage Workforce Transition Ladder that tracks how a role changes as automation deepens. It names six commitments a Caribbean board should make to its workforce before automation moves into production. And it ends, as the series always does, with five questions a director should ask — this time about the people, not the platform.

As in every edition, this paper is vendor-neutral and, here, methodology-neutral. Whether the enterprise redeploys, retrains, or reduces is a decision each board makes in its own context, with its own values and its own obligations. What this paper argues is narrower and firmer: that the decision belongs to the board, must be made in advance, and must be made in writing.

 

“The question is not whether AI will change Caribbean jobs. It is whether the change will be designed by the board that owns the enterprise — or discovered by the workforce that holds it together.” — Dr. Dawkins Brown, Executive Chairman, Dawgen Global

 

1. The change arrives as tasks, not headcount

In Edition 01 we argued that the most common cause of failed Caribbean AI pilots was not the technology but the imprecise definition of the use case. In this edition we make the workforce counterpart of that argument: the most common cause of botched AI workforce transitions is that boards think about them in the wrong unit. They think in headcount — how many roles will go — when they should think in tasks: which activities within a role are absorbed by the system, which are handed back to the human, and which are genuinely new.

A role is a bundle of tasks. Automation rarely removes the whole bundle. It removes some tasks, reshapes others, and — this is the part boards consistently miss — creates new ones: reviewing the system’s output, handling the exceptions it escalates, maintaining the data it depends on, and auditing the decisions it makes. The honest workforce question is therefore not “how many people does this replace?” It is “what does each affected role consist of after the system is live, and is that a role a person would want and the enterprise would value?”

This reframing matters because it changes who should be in the room. A headcount question is answered by finance and IT. A task question is answered by the people who actually do the work, the managers who supervise it, and the board that is accountable for both the enterprise’s capability and its people. The first discipline of a workforce transition is to insist on the task-level view before any number is attached to it.

 

The Task-Reallocation Map

Dawgen Global uses a simple map in Caribbean engagements to force the task-level view. Every activity in an affected role is sorted into one of four destinations — and the proportions tell the board, immediately, what kind of transition it is actually authorising.

Destination	What it means	The board’s real concern
Absorbed	The task is now done by the system, end to end. The human no longer performs it.	Does the institution lose a capability it may one day need to do manually — and who still knows how?
Reshaped	The task remains human but changes character — from doing to checking, from producing to judging exceptions.	Is this a richer role or a hollowed one? Reshaped work can elevate a person or reduce them to a rubber stamp.
Created	New tasks the system requires: oversight, exception handling, data stewardship, model assurance.	Who is trained and funded to do these — before the system is live, not after the first failure?
Removed	The task disappears entirely — it was only ever a workaround for the absence of the system.	Genuine capacity released. The opportunity is to redeploy it to higher-value work the enterprise has neglected.

 

Boardroom takeaway Ask for any AI proposal to be expressed as a Task-Reallocation Map for every affected role before any headcount figure is presented. A proposal that can only describe its workforce impact as a number — “saves 12 FTEs” — has not done the work. The number is a conclusion; the map is the analysis. Without the map, the board is approving an outcome it cannot see.

 

2. The Dawgen Global Workforce Transition Ladder

Edition 04 introduced an Authority Ladder for how much autonomy a board delegates to a system. This edition introduces its workforce companion: a four-stage ladder describing how a human role changes as automation deepens beneath it. Most Caribbean roles touched by AI will move up this ladder over time. The board’s job is to know which stage a role is at, decide whether it should move, and make sure the person in the role moves with it rather than being left behind by it.

Stage	What the human does	Board priority
S1 — Augmented	The person does the work; the system assists — drafting, summarising, surfacing. Output and accountability stay with the human.	Lowest disruption. Priority is fluency: making sure people can actually use the tool well.
S2 — Supervisory	The system does the work; the person reviews and approves before anything takes effect. The role shifts from producing to judging.	Reskilling. Reviewing is a different, harder skill than doing — and quietly demands more judgement, not less.
S3 — Exception	The system handles the routine autonomously; the person handles only what it escalates. Most of the old role is gone.	Role redesign and capacity. Fewer people are needed for the task — redeployment must be planned, not improvised.
S4 — Stewardship	The person no longer touches the work at all; they own the system that does — its performance, assurance, and improvement.	A genuinely new role. Few in number, high in value — and a deliberate career destination, not an accident.

 

Two cautions about the ladder. First, movement up it is not automatic progress for the person. A move from S1 to S2 can enrich a role or hollow it, depending on whether the human is genuinely exercising judgement or merely clicking “approve” on outputs they no longer understand. The board should treat the rubber-stamp risk as real: a supervisor who cannot meaningfully challenge the system is not a control, and is not in a richer job.

Second, the higher stages thin out. S3 and S4 need fewer people than S1 and S2. That is the point at which a transition stops being about reskilling and becomes about numbers — and it is the point Caribbean enterprises most often reach without a plan. The ladder’s value is that it lets a board see that point coming, sometimes years out, while there is still time to redeploy people into work the enterprise has been unable to staff.

 

“The danger is not the job that disappears. It is the job that quietly empties — the supervisor who approves what they can no longer challenge, the reviewer who has become a signature. A hollow role is a failed control and a failed person at once.” — The Dawgen Global Workforce Transition Principle

 

3. Six workforce commitments a board should make in advance

Edition 04 set out six governance additions for agentic deployments. This edition sets out the workforce parallel: six commitments a Caribbean board should make — in writing, before automation reaches production — to the people whose work it changes. As with the governance additions, they are not optional and not sequenced. They are simultaneous, and they are the board’s, not the project team’s.

3.1  Decide redeploy, retrain, or reduce — and say so

The board commits, before deployment, to which of the three paths each affected group is on, and communicates it honestly. Ambiguity is not kindness; it is the cruelty of letting people guess. People can plan around a hard truth told early. They cannot plan around a soft silence.

3.2  Fund the reskilling before the system, not after

Reskilling budgeted as an afterthought is reskilling that does not happen. The commitment is to fund the transition to S2/S3 skills — reviewing, exception-handling, data stewardship — as a line item in the automation business case itself, released before go-live, so the workforce is ready for the role rather than failing in it.

3.3  Protect institutional memory of the automated task

When a task is absorbed (S3/S4), the knowledge of how to do it manually decays. The commitment is to name who retains that capability and how it is kept alive — because the day the system fails, an enterprise that has forgotten how to do the work by hand has no fallback. This is a continuity control, not nostalgia.

3.4  Communicate before the rumour does

The commitment is that the workforce hears about a material change from leadership, in a planned way, before it leaks. In Caribbean enterprises — often close-knit, with long tenures and dense informal networks — a change discovered through rumour does more damage to trust than the change itself. Sequence and source are governance decisions.

3.5  Treat redeployment as the default, reduction as the considered exception

The commitment is a stated preference: released capacity is first offered to work the enterprise has been unable to staff — the backlog of higher-value activity every Caribbean firm carries. Reduction remains a legitimate option, but as a deliberate, documented decision, not the path of least resistance.

3.6  Assign a named executive owner for the workforce transition

As with the named system owner in Edition 04, someone must own this by name — typically the CHRO or COO — accountable to the board for the Task-Reallocation Map, the reskilling spend, the communication plan, and the redeployment outcomes. A transition that is everyone’s responsibility is no one’s.

 

How this connects to the Authority Ladder The two ladders move together. As a system climbs Edition 04’s Authority Ladder — from Read to Draft to Act — the human role beneath it climbs this edition’s Workforce Ladder, from Augmented toward Stewardship. A board that authorises more machine autonomy without planning the matching human transition has approved only half a decision. The governance of the system and the transition of the workforce are one authorisation, considered twice.

 

4. Where Caribbean enterprises should actually begin

As in earlier editions, the right first move depends on where the reader’s organisation already is. Three situations, three different starting points.

If you are still in wave one (a first pilot, mostly augmentation)

Your workforce risk is low today but your habits are forming now. Begin by building the Task-Reallocation Map into how you scope every pilot, so the task-level view becomes routine before the stakes rise. The discipline is cheap to adopt at S1 and expensive to retrofit at S3.

If you are scaling (wave two, moving roles toward supervisory)

This is the decisive moment, and the most commonly mishandled. Roles are shifting from doing to reviewing, reskilling is now urgent rather than theoretical, and the rubber-stamp risk is live. Make the six commitments now, fund 3.2 immediately, and name the owner in 3.6 this quarter. Most of the value — and most of the risk — of the whole transition sits in this band.

If you are already at exception or stewardship in some functions (wave three)

You are past the point where reskilling alone answers the question; you are into genuine capacity and numbers. Redeployment must now be actively managed against the enterprise’s real backlog of unstaffed higher-value work, and any reduction must be the documented, board-owned exception of 3.5 — not a drift. Institutional-memory protection (3.3) is no longer optional here; it is a continuity obligation.

 

5. Five questions a director should ask

As in every edition, the paper closes with five questions — here, the ones a director should ask before an automation programme moves into production. Each matters more in the workforce case than the technology case, because the technology can be rolled back and the trust cannot.

1. Has this been mapped at the task level — not just the headcount level?

If the only workforce figure in the proposal is a number of roles or FTEs, the analysis has not been done. Ask for the Task-Reallocation Map. The number is a conclusion; demand to see the reasoning beneath it.

2. For each affected group, have we decided redeploy, retrain, or reduce — and committed to telling them?

An undecided path is a decision deferred onto the workforce as anxiety. Ask which of the three each group is on, and when and how they will be told.

3. Is the reskilling funded and scheduled before go-live, in this business case?

Reskilling that depends on a future budget is reskilling that will not happen. Ask to see it as a released line item with dates, not an intention.

4. Who retains the ability to do the automated work by hand if the system fails?

An enterprise that automates a task and forgets how to do it manually has created a single point of failure with no fallback. Ask who, by name, keeps that capability alive.

5. Will our workforce hear this from us, in a planned way, before they hear it as a rumour?

In Caribbean enterprises especially, the sequence and source of the message shape trust more than the message. Ask to see the communication plan, and confirm leadership owns it.

 

The pre-mortem the board should run Before authorising production, run the workforce pre-mortem: assume it is eighteen months from now and the transition has damaged the enterprise. Two stories explain it. In the first, the system failed and no one remembered how to do the work — capability was lost. In the second, the people found out too late and stopped trusting leadership — the institution was lost a different way. The six commitments exist to make both stories impossible to tell. If the board cannot yet see why neither could happen, it is not ready to authorise production.

 

6. Where the series goes from here

With this edition, the series has moved through the full arc of a Caribbean enterprise’s AI realisation: the disciplined first pilot (01), the use cases worth funding (02), the move to wave two (03), the authority delegated to systems that act (04), and now the transition of the people whose work those systems change (05).

Two papers remain on the path Edition 04 pre-trailed. The next will take up the integrated, independent advisory model — what it means for a firm and a client to share an AI capability rather than the firm merely advising on a project. The final paper in this arc will consolidate the series into a complete Caribbean AI governance and workforce playbook: every framework, ladder, and checklist in one board-ready reference.

 

7. How Dawgen Global delivers the workforce transition

 

 

Dawgen Global works with Caribbean boards and executive teams to design and govern AI workforce transitions through our HR Advisory, Risk Management, and Business Advisory practices, supported by our curated network of global partners and vendors for the underlying technology. The engagement typically begins with Task-Reallocation Mapping across affected functions, proceeds to placing each role on the Workforce Transition Ladder, and produces the six written commitments and the named-owner mandate the board adopts before production.

Because the firm is independent and integrated, the same team that advises on the governance of the system (Edition 04) advises on the transition of the workforce (this edition) — so the two halves of the decision are made together rather than in separate rooms that never reconcile. That integration is the practical expression of the principle this paper argues: the system and the people are one authorisation, and one accountability.

Boards wishing to discuss an AI Workforce Transition Review for their enterprise can contact Dawgen Global at [email protected].

 

Dawgen Global  •  Big Firm Capabilities. Caribbean Understanding.

The Caribbean AI Realisation Series  •  Edition 05  •  “When the Work Changes Hands.”

About Dawgen Global