AI Training for Sri Lanka's Apparel Industry: Where the Hours Really Go
In Sri Lankan apparel, the fastest AI gains are in the office, not on the shop floor: merchandising, product development, planning, sourcing, compliance reporting and HR. These teams spend hours every week on buyer emails, comments, costing sheets, time-and-action calendars and audit paperwork, work that AI can speed up with a human checking every number. Start with merchandising and planning, keep buyer designs and pricing out of unapproved tools, and train supervisors in the language they work in.
The merchandiser at 9 pm
Anyone who has worked in Sri Lankan apparel knows the merchandiser who is still online at nine at night. The buyer is in Europe or the United States, the comments from the last fit session have just come in, the costing needs revising before morning, and three suppliers still have not confirmed fabric dates.
Most of that evening is not decision-making. It is reading, comparing, rewriting, chasing and updating. That is exactly the kind of work AI is good at speeding up, and it is why apparel offices are one of the clearest places in Sri Lanka to start an AI programme.
Why apparel, and why now
Apparel is Sri Lanka's largest manufacturing export. According to Joint Apparel Association Forum figures reported in January 2026, apparel and textile exports reached USD 5.02 billion in 2025, up 5.42% from USD 4.76 billion in 2024. The industry employs more than 350,000 people directly, by the JAAF chair's own figure, and up to a million more indirectly.
The industry competes on speed, quality, compliance and relationships with global buyers. Those buyers are adopting AI themselves, in their design, sourcing and planning teams. When the people on the other side of the email work faster, the suppliers who keep up keep the business.
And unlike many AI conversations, this one does not start with expensive systems. It starts with people who already have a laptop, an inbox and a spreadsheet, learning to use tools that are already in reach.
Where the hours go in the factory office
| Team | Repeated work | Where AI helps | Type |
|---|---|---|---|
| Merchandising | Buyer emails, fit and quality comments, order status updates, time-and-action calendars | Summarise long threads, turn comments into action lists, draft status updates, flag dates at risk | Acceleration |
| Product development | Comparing spec versions, sample requests, trim and fabric details | Compare two versions of a spec and list every change, draft sample request notes | Acceleration |
| Costing | Rebuilding cost sheets, explaining variances to buyers | Check formulas, explain a variance in plain language, prepare the buyer note | Augmentation |
| Planning | Line plans, capacity checks, weekly reports | Turn raw exports into a readable weekly report, spot gaps and clashes | Augmentation |
| Sourcing | Supplier follow-ups, quotation comparisons | Draft follow-ups, put quotations side by side, list questions to ask | Acceleration |
| Compliance | Audit preparation, buyer questionnaires, sustainability data | Organise evidence against each requirement, draft answers for review | Augmentation |
| HR | Recruitment messages, training material, policy explanations for large workforces | Plain-language versions in Sinhala and Tamil, onboarding question banks | Acceleration |
The type column uses our 3 As of AI. Most apparel teams should start in Acceleration, the same work done faster, before moving to Augmentation and, later, Automation of steps that repeat every week.
Merchandising: the biggest win
If an apparel company can only train one team first, it should usually be merchandising. Merchandisers sit at the centre of every order, they live in email and spreadsheets, and their working day is full of the reading and rewriting AI handles well.
The early wins are practical. A long thread with a buyer becomes a short summary with every open question listed. A set of fit comments becomes an action list by department, with owners and dates. A time-and-action calendar gets checked against the latest supplier confirmations, and the dates at risk are flagged before the buyer notices. A tricky email, the one that has to say no politely, gets drafted in a minute instead of twenty.
The rule that makes it safe is just as practical: AI drafts, the merchandiser decides. Every date, quantity and price in anything that goes to a buyer is checked against the source by a person.
Get the Sri Lanka AI Rollout Kit
Eight pages you can use this week: the 90-day plan on one page, a leadership readiness check, a PDPA-aware staff AI-use guideline, a training-needs survey, a use-case worksheet, a one-page business case and ten questions to ask any training provider.
Compliance and sustainability reporting
Buyers ask Sri Lankan manufacturers for more evidence every year: social compliance, chemical management, environmental data, traceability. Much of the effort goes into finding the right documents, matching them to each requirement, and writing clear answers.
AI is a strong assistant here, with one hard rule. It can organise, draft and check for gaps. It must never be allowed to fill a gap with something that sounds right. Every figure in a compliance or sustainability report has to come from a real record, and someone has to be able to show where. Training for compliance teams should spend as much time on checking as on drafting.
The shop floor is a different project
When people hear "AI in apparel" they often picture cameras spotting defects on the line, or machines predicting their own maintenance. Those are real, and manufacturers around the world are investing in them. But they are engineering projects with their own budgets and timelines.
Training is where the office side starts, and it pays back faster. It also builds the understanding leaders need before they sign off on bigger systems. A planning manager who uses AI every week asks much better questions of a vendor selling an AI scheduling system.
Buyer confidentiality and worker data
Apparel has two kinds of sensitive information, and staff need to know both.
Buyer information. Designs, tech packs, unreleased styles, prices and volumes usually sit under confidentiality terms with the buyer. They belong in the red category of any staff guideline: never in an unapproved AI tool, and only in approved tools when the company is sure that use fits its buyer agreements.
Worker information. Apparel companies hold personal data on very large workforces. With the main duties of Sri Lanka's Personal Data Protection Act due on 1 January 2027, HR teams in particular need clear rules before they use AI on anything involving employees. We set those rules out in the PDPA and AI at work.
A training plan for an apparel company
Start with a short leadership session, so plant heads and function heads agree on the outcomes. In apparel these are often things like faster responses to buyers, fewer late changes, or less time spent preparing for audits.
Then train merchandising and planning together. They share so much work that training them side by side builds shared prompts and templates from day one.
Then compliance, sourcing and HR, each on their own real tasks.
Name a champion in every plant, not only at head office. Plants outside Colombo are where support is hardest to reach, and a local champion who speaks the team's language makes the difference. Our post on training teams who work in Sinhala and Tamil covers how.
Measure before and after. Pick two or three tasks, such as turning around buyer comments or preparing the weekly order status, and time them before training and again at the end. That is your business case for the next phase. The full sequence is in our 90-day rollout plan, and if your company is part of a wider group, see training a whole group on AI.
One thing to try this week: ask three merchandisers what kept them online last night. Write the answers down. That list is your first AI training brief.
Get the Sri Lanka AI Rollout Kit
Eight pages you can use this week: the 90-day plan on one page, a leadership readiness check, a PDPA-aware staff AI-use guideline, a training-needs survey, a use-case worksheet, a one-page business case and ten questions to ask any training provider.
- How to Roll Out AI Across a Sri Lankan Company in 90 Days
- Sri Lanka's PDPA and AI at Work: What Staff Can and Cannot Paste into ChatGPT
- What Does Corporate AI Training Cost in Sri Lanka? How to Compare Quotes
- AI Training for Teams Who Work in Sinhala and Tamil
- AI Training for Sri Lanka's Apparel Industry: Where the Hours Really Go
- AI Training for Sri Lankan Hotels and Tourism Teams
- After the Brain Drain: How AI Upskilling Helps Sri Lankan Teams Do More With the People Who Stayed
- Training a Whole Group on AI: A Playbook for Sri Lanka's Diversified Conglomerates
Frequently asked questions
How can AI help Sri Lanka's apparel industry?
The quickest gains are in office functions: summarising buyer emails and comments for merchandisers, comparing spec versions, explaining costing variances, turning planning data into weekly reports, organising compliance evidence and producing plain-language HR material in Sinhala and Tamil. Shop-floor systems such as defect detection are separate engineering projects.
Which apparel team should be trained on AI first?
Usually merchandising, often alongside planning. Merchandisers handle large volumes of email, comments and spreadsheets, which is the kind of work AI speeds up most, and every improvement there affects every order.
Is it safe to put tech packs or buyer prices into ChatGPT?
Treat designs, tech packs, unreleased styles, prices and volumes as confidential buyer information. Keep them out of unapproved AI tools, and only use approved business tools where the company has confirmed that use fits its buyer agreements.
How big is Sri Lanka's apparel industry?
Apparel and textile exports were USD 5.02 billion in 2025, up from USD 4.76 billion in 2024, according to JAAF figures reported in January 2026. The industry employs more than 350,000 people directly.
Can AI write sustainability or compliance reports for apparel buyers?
AI can organise evidence, draft answers and find gaps, but every figure must come from a real record and be checked by a person. It should never be used to fill a missing data point with something that only sounds right.
Sources
- Ada Derana, apparel and textile export earnings for 2025 (JAAF data), January 2026
- Daily Mirror, "Apparel sector upbeat on prospects to achieve US$ 8bn export target" (employment figures)
Start with the people closest to the buyer. Map your merchandising and planning teams for free.
Map your team for free →
