Client accounts of Kilau Neural engagements

What firms say about
working with us.

These are accounts from businesses that engaged Kilau Neural for ML capability builds, evaluation frameworks, and AI briefings. We present them without editing for sentiment.

Back to Home
47
4.7
8
12

Direct accounts from past clients

ZA

Zulaikha Ahmad

"We brought them in for an Evaluation Framework engagement after realising our team was shipping model updates with no reliable way to measure whether the changes helped or hurt. The test set they built from our production data exposed several problems we had not noticed. The written framework is now part of our standard deployment checklist."

April 2025 — Evaluation Framework

RL

Raymond Lim

"I attended the Senior Briefing with three members of our board ahead of a significant AI procurement decision. The session was well-prepared and direct — they covered data residency and governance obligations in a way our board could actually use. The written note the next morning was exactly what we needed to circulate internally. Would have preferred more time on the financial services regulatory specifics, but overall very useful."

March 2025 — Senior Briefing

NI

Nurul Izzati

"The Capability Build engagement took about three and a half months in the end. What stands out is that the documentation they left behind is actually usable — our engineering lead was able to trace every design decision back to a written rationale. Six months on, we have extended the system ourselves without needing to contact them."

January 2025 — Capability Build

CT

Chen Teck Wei

"They declined two of the three problems we brought to them at scoping, on the grounds that the data we had would not support what we were hoping to achieve. That kind of honesty is rare. The one problem they agreed to take on was delivered within scope and on time."

February 2025 — Capability Build

SR

Siti Rohana

"We engaged them for the evaluation framework after a near-miss where a model update degraded badly in production and we had no systematic way to catch it before it reached customers. The work was solid and the process well-documented. The engagement ran slightly longer than planned because our team was slower to provide production samples than expected — that was our side, not theirs."

April 2025 — Evaluation Framework

AM

Ahmad Mukhriz

"The briefing format was well-suited to our situation. We had board members who were sceptical of the cost and risk of an AI initiative, and members who were pushing to move faster than our data infrastructure could support. The session gave everyone a shared factual basis for the conversation. The follow-up note was clear enough to include in board papers."

March 2025 — Senior Briefing

Engagement accounts in detail

Demand forecasting for a Malaysian consumer goods distributor

Duration: 14 weeks

The firm's purchasing team was using spreadsheet-based forecasting that was calibrated to historical averages and failed to account for seasonal variation across their 300+ SKU range. Overstock on slow-moving lines and stockouts on fast-moving ones were both frequent and costly.

A time-series forecasting system trained on three years of sales data, segmented by product category and distribution region. The pipeline ran weekly from their existing ERP export, required no additional tooling, and produced outputs in a format their purchasing team could use directly.

Stockout frequency reduced by 34% in the first four months post-deployment. Overstock holdings fell by 19%. The purchasing team trained on the system in two sessions and has since run it independently for over eight months without external support.

"The scope was agreed clearly enough that there were no surprises on either side. When our data turned out to be messier than expected, they adjusted the pipeline and documented the changes without needing to renegotiate."

Classification system evaluation for a KL fintech firm

Duration: 7 weeks

The team had a working document classification system in production but was deploying updates based on internal accuracy scores that were not representative of the production data distribution. A regression had gone undetected for two months before a client complaint surfaced it.

A production-representative test set of 1,400 labelled examples, stratified by document type and time period. A written evaluation protocol covering five metrics relevant to their deployment constraints. Integration with their existing deployment pipeline so that each release runs against the test set automatically.

The evaluation pipeline caught a model regression in the first release cycle after implementation. The team's confidence in deployment decisions improved visibly — release frequency increased while customer-facing error rates fell by roughly 40% over the following quarter.

"What we had before was an evaluation process in name only. After the engagement we had one that actually told us something. The production test set alone was worth the cost."

Board-level AI briefing for a Malaysian logistics group

Duration: 1 day

A logistics group with operations across West Malaysia was evaluating whether to procure a vendor AI solution or commission internal development. The board had differing views on the risk and cost of each option and lacked a shared factual basis for the discussion.

Build vs buy trade-offs specific to their use case. Data residency obligations under PDPA for the vendor hosting options they were considering. Governance and disclosure requirements that would apply under the proposed ASEAN AI governance framework. Internal capability requirements for each procurement path.

The board reached a unanimous decision at the next board meeting to pursue a scoped internal build rather than the vendor solution, citing data residency concerns and total cost of ownership over five years. The follow-up note was appended to the board paper.

"They presented the options without advocating for one. That was exactly what we needed — a structured way to think through the decision, not a recommendation that happened to align with their commercial interests."

Reach us directly

+60 3-2074 5198
53, Jalan Hang Lekiu
50100 Kuala Lumpur
Wilayah Persekutuan

Mon–Fri: 9:00 am – 6:00 pm
Sat: 10:00 am – 1:00 pm

Practice credentials

MDEC Digital Transformation Partner

Listed in the Malaysia Digital Economy Corporation's AI practice partner network.

ASEAN AI Practitioners Forum — 2024 Recognition

Recognised for responsible AI deployment practice across Southeast Asian enterprises.

MIA Technology Advisory Panel

Contributing member providing input on AI governance for accounting and finance professionals.

Add your organisation to the list.

We respond to every enquiry within two working days. No obligation, no extended sales process. Just a brief note on whether and how we can help.

Send an Enquiry