Client feedback Kelulut Labs

What clients found useful — and what they'd note for others

← Back to Home

47

Clients served

4.7

Average satisfaction (out of 5)

94%

On-time delivery rate

3

Years in practice

What clients have said

LH

Lim Hui Shan

Trading company, Georgetown

The Data Tidy-Up was worth it just for the change note. We'd been carrying over the same inconsistencies for years — three different spellings of the same supplier name, rows that had been duplicated when someone copied the sheet. Now we have a clean file and a record of why each change was made, which helped us explain it to the accounts team.

June 2025 · Data Tidy-Up

RM

Razif Mahmud

Electrical supplies, Butterworth

The Quotation Drafting Helper has cut the time we spend on each quote by roughly half. The first draft now has our payment terms and contact details already in place — we just fill in the specific items and amounts. I'd say the handover session could have been longer, maybe another hour to cover more edge cases, but the follow-up email option helped with the questions that came up later.

June 2025 · Quotation Drafting Helper

TW

Tan Wei Leng

Café chain, Penang

The Team Session was not what I expected. I thought it would be a product demonstration. Instead it was three hours of our staff asking direct questions about how to spot when an AI tool is making something up, and the trainer actually answering with examples from our own supplier forms. The checklist is still on the wall in the back office.

July 2025 · Team Session

SA

Siti Aminah Osman

HR consultancy, Georgetown

Booked the Data Tidy-Up because we were about to switch to a new system and wanted to start clean. What I valued most was that sensitive employee data was returned in a separate file rather than put through the same cleaning process as the general records. That distinction mattered for us.

July 2025 · Data Tidy-Up

KP

Kumar Perumal

Auto parts retailer, Ipoh

The Team Session was delivered over video because my team is split between two locations. Worked reasonably well. The trainer used our own invoice templates as the examples, which kept the discussion grounded. Staff appreciated that it wasn't just a general intro — they asked more questions than I expected.

June 2025 · Team Session

YF

Yee Fong Chai

Food distributor, Penang

We'd looked at a few AI consultancies before this one. Most wanted a discovery engagement and a proposal. Kelulut Labs listed the prices and what each service covers on the website. We picked the Quotation Helper, paid the fixed fee, got the template. It does what it says it does. That's what we needed.

July 2025 · Quotation Drafting Helper

Three engagements in detail

Case Study 01 · Data Tidy-Up

A hardware distributor's product catalogue

Challenge

A Penang hardware distributor had a product catalogue that had grown to 4,200 rows over eight years. Multiple staff had added to it using different naming conventions — some categories in English, some in Malay, some abbreviated. The same product sometimes appeared under three different names. The company wanted to use AI to help with inventory queries but found that inconsistent naming produced unreliable answers.

What We Did

We reviewed the catalogue and settled on a single naming convention in agreement with the client — English names for product categories, with Malay alternatives recorded in a separate column. Duplicate entries were identified and merged. Sensitive supplier pricing columns were returned in a separate sheet rather than included in the cleaned main file. The process took five working days.

Outcome

The cleaned catalogue had 3,810 rows — 390 duplicates removed. The client reported that inventory queries run against the cleaned file now return consistent results. The change note identified seven product categories that had been split across two naming conventions, giving the team a reference for future additions.

"The change note was the thing we didn't know we needed until we had it."

Case Study 02 · Quotation Drafting Helper

A building materials supplier's quotation process

Challenge

A Kedah building materials supplier sent between twelve and twenty quotations a week. Each was typed from scratch by the same staff member — same header, same payment terms, same validity clause — because there was no standard template. The person handling quotations estimated the typing alone took thirty to forty minutes per quote.

What We Did

We read eighteen of the client's existing quotations, identified the standard structure and recurring terms, and built a template library that pulled from the product spreadsheet the client already maintained. The system produced a draft with the correct header, payment terms, and validity period — the staff member then fills in quantities and line items. Handover session covered correction of drafts that didn't match a non-standard situation.

Outcome

The client reported that time spent per quotation dropped from thirty–forty minutes to around twelve minutes. The staff member doing quotations described the change as removing the part of the task they found most repetitive. All quotes still go through the same person before sending; the human review step was not changed.

"It sounds like our quotes now. That was the important thing."

Case Study 03 · Team Session

A professional services firm's staff awareness session

Challenge

A Georgetown accounting firm had six staff using AI chat tools to help with document drafting. The partners had no clear view of what was being pasted in or what the staff were relying on the tools to produce without checking. The concern was less about efficiency and more about confidentiality and accuracy in client-facing documents.

What We Did

Before the session, we asked the firm for samples of the document types staff used — engagement letters, client summaries, internal memos. The session covered what types of content should not be pasted into external AI tools, how to verify whether a generated draft had introduced inaccurate figures, and what a practical review process for AI-assisted documents might look like. The session was held at their office with seven staff attending.

Outcome

The firm established a simple internal rule for which document types required a second-person review when AI had been used in drafting. One senior staff member described the session as the first time they'd had a conversation about this that wasn't hypothetical. The printed checklist was adopted as part of the firm's document review procedure.

"We didn't stop using these tools. We just got clearer on where to check."

Reach us directly

Phone

+60 4 261 7398

Mon–Fri 9am–6pm, Sat 9am–1pm

Email

[email protected]

Typically responded to within one working day

Address

68 Lebuh Pantai
10300 George Town
Pulau Pinang

Office Hours

Monday – Friday: 9:00 am – 6:00 pm
Saturday: 9:00 am – 1:00 pm
Sunday & Public Holidays: Closed

Professional standing

MDEC Digital Transformation Partner

Registered with Malaysia Digital Economy Corporation for small-business digital services.

Penang SME Association — Associate Member

Active member since 2023, participating in local SME development discussions.

Describe your situation and we'll take it from there

An initial message doesn't bind you to anything. We'll tell you which of the three services applies, or suggest a conversation if your case is less straightforward.

Get in Touch