On 21 September 2026, Reuters reported forthcoming Inter-American Development Bank research suggesting that AI could expand output in Latin America and the Caribbean while leaving workers with very different wage outcomes. The distinction depends partly on whether people can move into activities that grow.

For business leaders, that is a reason to examine the transition between jobs and tasks, not simply the speed of software adoption. Training has limited value if there is no credible route into work where the new capability can be used.

What happened

Reuters reported IDB scenarios in which broad adoption could leave the regional economy 5.1% larger after a decade. Wages could rise where workers move into expanding sectors, or fall by as much as 20.9% where that transition does not happen. The full report was expected in November.

These are conditional estimates for a particular region, not observed outcomes or a forecast for an individual employer. The reporting does not justify transferring the figures to Dubai, treating them as a budget assumption or promising that training alone will deliver the favourable scenario.

Why it matters

An organisation introducing AI should be able to explain where displaced effort is expected to go. ‘People will do higher-value work’ is an intention until someone identifies that work, its demand and the preparation needed to perform it.

Consider a hypothetical operations team whose routine document preparation becomes faster. The company wants employees to spend more time resolving complex customer cases. That transition requires more than an available hour: staff need access to evidence, authority to make decisions, supervised practice and someone able to review the difficult cases.

There may also be fewer suitable opportunities than the proposal assumes. An honest workforce plan should recognise that possibility early. It should compare the amount and type of work expected to grow with the people whose current activities may shrink.

Managers have a role in making movement possible. A receiving team needs capacity to support learners. An existing manager may need to release a capable employee rather than retain them to protect a local target. Those choices belong in the change plan alongside the technology milestones.

The bigger shift

The relevant competition may include how effectively organisations move knowledge and people between activities. Buying access to AI is one decision; building a route from an old responsibility to a useful new one is another.

A practical transition plan names the destination task, the competence required, the learning opportunity and the evidence that the employee can perform it. It also states what happens if the expected demand does not materialise. That gives employees something more concrete than reassurance about an unspecified future role.

The argument in Why AI Strategy Beats AI Tools is relevant here. The organisation needs a purpose for the capability it acquires. Without that direction, additional training can become another activity completed without a clear operational result.

My take

I would ask for a workforce transition plan beside the AI investment proposal. It should show which responsibilities may decline, which could expand and what must happen for people to move between them.

Review that plan with employees who understand the work and with the managers expected to receive the new capacity. Check whether the proposed destination has real demand, a suitable learning route and resources for supervision.

Measure the transition itself: can people perform the revised work, are opportunities accessible and is the receiving team equipped to support them? Course attendance and tool logins cannot answer those questions.

The IDB figures should prompt examination, not numerical imitation. A company’s next step is to test its own assumptions about changing work and make the path between responsibilities visible before promising the benefits of adoption.