This is the second in a series on repositioning Jamaica’s global services sector. The first column argued that the sector’s contraction was a productivity failure, not an Artificial Inteligence displacement, and that AI is the most direct fix available.

Jamaica no longer has a cost advantage in outsourcing. The 2026 benchmarks put us at parity with Colombia and Mexico and nowhere near Asia. That is not a crisis. It is permission to stop fighting a war we were always going to lose — and to start winning the one that actually pays.

In my last column, I argued that Jamaica did not lose 20,000 outsourcing jobs to AI — we lost them to productivity, and AI is how we get them back. I want to continue that thought, because the productivity conversation leads somewhere uncomfortable that we need to sit with as an industry.

If AI lifts output per agent hour, the obvious question is: output toward what? And the honest answer is that we have been pointing our productivity at the wrong prize for a decade.

Look at what we cost

Here are the 2026 loaded per-hour benchmarks the buyers are actually working from:

•Jamaica and Trinidad: roughly US$12 to US$18

•Colombia and Mexico: US$12 to US$18

•The Philippines: US$8 to US$12

•India: US$6 to US$9

Read that again slowly, because it rewrites our entire pitch. We are no longer cheaper than our nearshore rivals — we are level with them. And against the offshore giants we are, and always will be, expensive. A client chasing the lowest sticker rate on transactional voice was never going to land in Montego Bay, and no incentive package we can afford will change that arithmetic.

The maths has moved. Asia owns the bottom of the market permanently, and Latin America has drawn even with us on price while building talent pools many times the size of ours.

So let me say the quiet part plainly: We cannot win on price, we should stop trying, and the sooner we internalise that the sooner we can compete where we actually have an edge.

The edge we keep underselling

Now, look at the other thing those same 2026 benchmarks say. Caribbean nearshore delivers the strongest customer-satisfaction-to-cost ratio in the entire nearshore field for US-time-zone, native-English voice work. When analysts describe where Jamaica specifically fits, the language is consistent: for English-language operations where accent clarity and cultural alignment matter most, the Caribbean is the strongest fit on the board.

That is not a small thing. That is the whole game — if we play it. Native English at genuine fluency. A US and Canadian time-zone overlap that lets a client fly in for a morning meeting and be home for dinner. A service instinct that shows up as empathy on a hard call, not as a script read at speed.

Those attributes are worthless on a password-reset queue that a bot will own by next year. But they are decisive on the work that is growing fastest and paying the most: complex technical support, healthcare, financial services, collections, trust-and-safety, and the judgement-heavy processing that the industry now calls knowledge process outsourcing — underwriting, claims adjudication, legal review, financial analytics.

That is regulated, high-empathy, high-stakes work. It is precisely the work that cannot be handed to the cheapest offshore seat, because a 15-cent saving means nothing next to a mishandled patient call or a compliance breach. And it pays two and three times what basic voice pays.

Why this is only possible now

For years the honest objection to “move Jamaica up the value chain” was capacity. Complex work needs experienced agents, experienced agents take a long time to develop, and we are a small country with a tight labour market and long ramp times. You cannot staff a health-care programme with people you are still training. That is exactly the constraint AI dissolves — which is why these two arguments are one argument.

The best-evidenced effect of AI in this industry is that it compresses the distance between a new hire and a competent one. Real-time agent assist puts the right procedure, the right disclosure, the right next step in front of the agent at the moment of the call.

Source-grounded knowledge retrieval means a six-week agent has the same reference depth as a six-year veteran. Automated after-call work and AI quality assurance free supervisors from busywork to actually coach. The productivity technology I wrote about last time is the same technology that lets a small, young workforce credibly deliver complex, regulated work it could not have touched three years ago.

In other words: AI does not just make our existing work cheaper. It makes better work reachable. Those are different claims, and the second one is the one that rebuilds the sector.

There is a comfortable version of the AI story we deploy the tools, run the same transactional programmes with fewer people, book the savings, and call it modernisation. That path ends with a smaller, cheaper, more fragile sector that is still losing the price war one quarter at a time — just with better dashboards.

The tools are neutral. They will cut cost on low-value work or unlock high-value work depending entirely on where we point them. If we use AI only to defend the bottom of the market, we are automating our way to the exit. If we use it to reach the top, we are building something durable. Same technology, opposite destinies. The choice is ours and it is a choice of ambition, not of budget.

What it asks of each of us

Operators have to change what we sell and how we price it. Stop quoting seats and start quoting outcomes — resolved cases, retained customers, clean claims. Build the certifications and the security posture that regulated work demands, because you cannot bid health-care or financial services on hospitality alone. And point every AI investment at moving people up, not merely thinning them out.

Government has to market the country we are becoming, not the one we were. Our competitors never stopped selling, and they are selling on cost we can no longer match. Our national pitch has to lead with the thing that is actually true and actually scarce: the best English-language, US-time-zone, high-CSAT delivery in the nearshore world.

The incentive regime should reward complexity and value retained on-island, not headcount for its own sake. And the enabling infrastructure — power, telecoms, disaster continuity — has to be treated as export infrastructure, because Melissa proved it is priced into every deal we lose.

The talent pipeline has to be rebuilt for the job as it is becoming. If AI owns tier one, the entry-level role is no longer script-reading — it is exception handling, judgement, empathy, and supervising machine output. That is a different curriculum, and HEART/NSTA Trust, the universities, and industry need to be designing it together, starting now.

The bottom line

The price war is over and we did not win it — and that is the best news this sector has had in two years, because it frees us to stop pretending. We were never going to be the cheapest. Chasing “cheapest” is what left us competing on a number for which Asia will always beat us, and Latin America has now caught us.

What we can be is the place North American companies send the work that is too important, too regulated, or too human to trust to the lowest bidder. AI is what makes that move possible for a country our size. Ambition is what makes it happen.

We lost the race to the bottom. Good! Let’s go win the one that’s worth running.

Yoni Epstein is president of Global Services Association of Jamaica as well as founder and chairman of itel.