After the Brain Drain: How AI Upskilling Helps Sri Lankan Teams Do More With the People Who Stayed

By Ammar Ahamed, Founder, Cocoon · 21 October 2026 · Part 7 of 8 in the Corporate AI in Sri Lanka series

The farewell cake

Most Sri Lankan offices know the routine by now. A WhatsApp message goes round, someone orders a cake, there is a short speech in the meeting room, and a colleague who knew how everything worked flies out the following week to Canada, Australia, the Gulf or New Zealand.

The cake is the easy part. The hard part comes a month later, when nobody can remember how the year-end report was put together, a key client asks for "the person who used to handle this", and the people who stayed are covering two jobs.

We cannot fix migration with training. But we can help the teams who stayed carry the load better, keep more of what leavers knew, and make staying a better choice.

What the numbers say

The Central Bank of Sri Lanka publishes departures for foreign employment from Sri Lanka Bureau of Foreign Employment data. The last four years look like this:

Under the Bureau's newer classification, about 77.7% of departures in 2024 and 76.2% in 2025 were classed as skilled. The Bureau no longer publishes a separate series for professionals, so we cannot say exactly how many were accountants, engineers, nurses or IT staff. But anyone running a team in Colombo knows the pattern from experience.

What leaving takes with it

When an experienced person leaves, three things go with them.

Capacity. Their work does not disappear. It lands on the people who stayed, often without a pay rise, and sometimes for months while a replacement is found and trained.

Knowledge. The most valuable things people know are rarely written down: which client needs a call before an invoice, why the costing sheet has that odd extra column, which supplier always ships late in December. That kind of knowledge leaves on the same flight.

Morale. Every farewell makes the people who stayed ask whether they should go too. A team that feels stretched and stuck loses its next person faster.

AI can help with all three, if it is introduced the right way.

1. Capacity: give the routine work to AI

The quickest relief for a stretched team is taking routine reading and writing off their plate: summarising long email threads, drafting standard replies, turning raw data into a first-draft report, preparing meeting notes, checking documents for missing information.

This is what we call Acceleration in our 3 As of AI: the same work, done faster, with a person still holding the pen. It is the safest place to start, because the person doing the work can see and check every output. It is also where a stretched team feels the difference first.

The key is training people on their own real tasks, not on generic examples. A team covering for two missing colleagues does not have time for a theory session. They need to leave with something that saves them time next week.

2. Knowledge: capture it before it walks out

This is the use most companies miss, and in Sri Lanka right now it may be the most valuable.

When someone resigns, there are usually a few weeks before they go. AI makes it possible to turn those weeks into a proper handover, not a rushed one. Here is a simple practice any team can run.

Map the process first. Sit with the leaver and map each recurring task they own: the steps, the systems, the people involved, the exceptions. Our free Workflow Mapper turns a messy description into a clear flowchart.

Record the walkthroughs. With their agreement, record the leaver explaining each task as they do it. A transcription tool, such as our notetaker, turns that into text.

Turn it into a guide. Use AI to draft a step-by-step guide from the transcript and the map. Then the leaver checks and corrects it, which is the most important step.

Test it with the successor. Have the person taking over follow the guide while the leaver is still there to answer questions. Fix whatever confused them.

Over time, those guides can become a team knowledge assistant that answers "how do we do this?" from your own documents. Our RAG Lab shows, hands-on, how that kind of assistant works and where it goes wrong.

3. Retention: a reason to grow here

People rarely leave for salary alone. They also leave because they cannot see themselves growing where they are. Skills are part of the answer to that.

An employer that teaches its people to build with AI, and gives them real problems to apply it to, is offering something many people would otherwise go abroad to find. It does not stop everyone leaving. But it changes the conversation from "what is there for me here?" to "this is where I am learning the most".

The AI champion role is a good example. Choosing someone curious in each team, training them more deeply, and giving them visible responsibility for helping colleagues is a growth path that costs little and means a lot. Our AI Champion programme is built around exactly that role.

There is a fairness point here too. If AI training only goes to head office, the teams outside Colombo, often the most stretched, are left out of the very thing that would help them most.

What AI is not for

It is tempting, especially for a finance team looking at a hiring freeze, to see AI as a way to avoid replacing the people who left. We would push back on that.

Stretched teams can tell when "AI" is code for "you will be doing three jobs now". If that is how the programme is introduced, people will resist it, or worse, the best of them will add it to their reasons to leave. AI used well gives people time back and lets them do better work. Used as a reason not to hire, it burns out the people you most need to keep.

So say plainly what the programme is for: to make the work lighter and the team stronger. Then make sure the first results prove it.

A plan for HR

Find the teams under most strain. Look at recent resignations and open roles by team. Those teams should go first, not last.

Start a handover practice now. Make the four-step knowledge capture part of every resignation, starting with the next one.

Train on real work, in small groups. Each person brings the task that costs them most time. Everyone leaves with something that works.

Name champions and recognise them. Make the role visible, give it time, and mention it in reviews.

Measure honestly. Time saved on tracked tasks, readiness before and after, and over a longer period, whether attrition in trained teams changes. Our 90-day rollout plan shows how to put it all in one quarter.

One thing to try this week: think of the last person who left your team. Write down three things they knew that nobody else does now. Then decide who is next most likely to leave, and start their handover guide while there is still time.

Frequently asked questions

Sources

1. Central Bank of Sri Lanka, Workers' Remittances and Labour Migration Bulletin 2025 Q4
2. Central Bank of Sri Lanka, Workers' Remittances and Labour Migration Bulletin 2024 Q4
3. Central Bank of Sri Lanka, Workers' Remittances and Labour Migration Bulletin 2023 Q4