George Town, Penang

A small firm that takes a measured view of AI

Our Story

Why Kelulut Labs exists

Kelulut Labs started from a straightforward observation: most small businesses in Malaysia were being told that AI would transform their operations, but very few were being given anything concrete to do on Monday morning. The conversation was loud on capability and quiet on scope.

We set up at Lebuh Pantai in George Town to work with businesses that wanted a specific kind of help — not a sales pitch, not a subscription, but a defined piece of work with a beginning and an end. The three services we offer came directly from conversations with local shop owners, trading companies, and small professional firms about what was actually slowing them down: messy records, time spent typing quotations, and staff who weren't sure what these tools were or weren't safe to use for.

The name comes from the stingless bee, kelulut — a species native to this region that builds structured, orderly comb. We liked the idea of small, well-organised cells of work. Nothing dramatic, just careful structure.

We don't have a house view on which AI platform is best. We don't sell software. What we do is look at what a business already has, decide what's worth tidying, and make one specific thing work better. Then we hand it back.

The Team

People behind the work

Ahmad Kamal

Founder & Data Lead

Spent twelve years managing inventory and procurement records for a George Town trading company before turning that experience into a methodology for cleaning small-business data. He handles every Data Tidy-Up engagement.

Siew Wan

Workflow Specialist

Background in operations for a small Penang export firm. She built the Quotation Drafting Helper methodology after watching colleagues retype the same terms onto every new quote for years. Leads all template-build engagements.

Rajan Nair

Training Lead

Ran staff development programmes for a regional logistics firm for eight years. He designs and delivers the Team Session, adapting examples to each client's own documents so the training is immediately relevant rather than hypothetical.

How We Work

Standards we hold ourselves to

Scope defined in writing first

Before any work begins, we confirm in writing what the engagement covers, what it doesn't, and what you'll receive when it's done. Nothing starts on a handshake with details to follow.

Your files stay yours

We don't retain copies of client data after an engagement closes. Sensitive rows flagged during Data Tidy-Up are returned separately rather than processed through external services.

Deliverables, not promises

Each service produces something tangible: cleaned files and a change note, a working template connected to your spreadsheet, or a session with slides and a printed checklist. The output is specified before you pay.

One person responsible per engagement

The team member leading your engagement is named at the start. They're your single point of contact throughout, which means you don't repeat context to a different person each week.

We say when something won't work

If an enquiry describes a situation where none of our services would help, we say so rather than find a way to fit it. It's a policy that costs us short-term but tends to produce longer-term working relationships.

Follow-up included, not billed separately

The Quotation Drafting Helper includes a short handover session for two staff. The Team Session includes a follow-up email address for questions over the next month. These are part of the price listed, not extras.

Our Approach

Working with AI tools in a small-business context

The businesses we work with in Penang and across Malaysia typically have between two and thirty staff. Their records have often grown organically — a spreadsheet started years ago by one person, inherited by another, and now used by three. These files can be used for AI-assisted work, but only after someone has gone through them carefully. That's what Data Tidy-Up addresses.

Quotation drafting is time-consuming not because it's complex but because the same information gets typed afresh each time. A tool trained on your own past quotes doesn't introduce a new system — it shortens an existing step while keeping the human review that clients expect. The person approving and sending the quote remains the same; only the time spent on the first draft changes.

Staff training on AI tools works differently when it focuses on scepticism rather than adoption. The Team Session is designed around the question of what to check rather than what to trust. Staff who understand how these tools produce outputs — and where that process fails quietly — are better positioned to use them well than staff who've been told they're powerful and left to figure out the limits themselves.

Kelulut Labs operates from George Town because the local small-business community is dense and varied enough to support this kind of focused, practical work. We're available for in-person engagements in Penang and for remote work elsewhere in Malaysia.

Work with us

Describe your situation and we'll suggest a starting point

An initial message doesn't commit you to anything. We'll tell you which service fits, or whether a conversation would be more useful first.

Get in Touch