45+
ENGAGEMENTS
4.8
AVG. SATISFACTION
5+
YEARS OPERATING
92%
ON-TIME DELIVERY
CLIENT REVIEWS
"We'd tried to build an internal anomaly flagging system twice before calling Cerebrix. Both attempts ended with something that worked in testing and broke in production. What they delivered was narrower in scope but actually ran reliably on our transaction data from day one. The documentation meant our team could manage threshold adjustments without going back to them."
Ahmad Hafiz
Head of Risk Technology, Kuala Lumpur
March 2026
"Our case library goes back over fifteen years and most of it was unstructured text. The knowledge graph they built let associates query across case precedents and internal memos in a way that simply wasn't possible before. The ontology design process took longer than we expected — there were a few rounds of stakeholder review — but the end result was worth it."
Siti Nabilah Tan
Senior Partner, Legal Practice, KL
February 2026
"We commissioned the AI ethics review ahead of a consumer credit scoring rollout. The scorecard identified three areas we hadn't fully considered — one of them would have created a disparity in approval rates across income bands. We delayed the rollout by three weeks to address it. That conversation with the regulator went much better as a result."
Raj Kumar
Chief Data Officer, Financial Services
January 2026
"The pricing and scope clarity made the decision straightforward to bring to our board. A fixed MYR 2,200 for the ethics review with a clear deliverable — no vague day-rate that could balloon. The scorecard itself was specific and actionable, not a generic checklist reworded for our context."
Faridah Yusoff
VP Operations, Insurance Group, KL
March 2026
"We have sensor networks across three facilities and were getting flooded with alerts. The calibration work Cerebrix did reduced noise by a substantial margin while actually catching a cooling anomaly that would have escalated into a shutdown. The threshold logic is now something our team adjusts directly when conditions change."
Lim Wei Jian
Infrastructure Manager, Selangor
February 2026
"We started with the ethics review as a way to assess whether Cerebrix was right for a larger anomaly detection project. It was a sensible way to see how they work — structured, honest about what the findings meant, and no overselling of what a remediation would require. We've since engaged them for the detection system and the experience has been consistent."
Nurul Zahirah
Head of AI Governance, Research Org
March 2026
CASE STUDIES
CASE STUDY 01 — ANOMALY DETECTION · FINTECH · KL
Challenge
A digital payments operator was handling high transaction volumes with a rule-based alert system that produced a large number of false positives each day. Staff were spending significant time reviewing alerts that resolved as legitimate, reducing the capacity to investigate the ones that mattered.
Solution
Cerebrix built a baseline behavioural model from six months of historical transaction data, distinguishing normal variation from structural anomalies. Thresholds were calibrated against the existing rule set, and the alert routing was connected to the client's incident platform with severity tiering.
Results
False positive rate dropped by around 68% in the first four weeks. The team shifted from reviewing every alert to reviewing only those flagged at medium severity or above. Three genuine fraud patterns were caught in the first two months that would likely have been missed in the volume of the old system.
Delivered in 7 weeks · MYR 5,100
CASE STUDY 02 — KNOWLEDGE GRAPH · LEGAL FIRM · KUALA LUMPUR
Challenge
A mid-sized legal practice had accumulated over 18,000 documents across case files, legal opinions, and internal memos. Finding relevant precedent required manual search across systems that didn't talk to each other. Knowledge was effectively siloed by individual practitioners.
Solution
Cerebrix designed a legal ontology covering cases, statutes, parties, outcomes, and associated matters. Entity extraction pipelines were run across the document archive, with manual review on a sample set to validate accuracy. A query interface allowed associates to search by entity type, date range, and relationship.
Results
Time spent on precedent research reduced substantially. Associates reported finding relevant cases they hadn't been aware of from earlier years. The graph was subsequently extended by internal staff using the schema documentation — no Cerebrix involvement required.
Delivered in 11 weeks · MYR 7,300
CASE STUDY 03 — AI ETHICS REVIEW · FINANCIAL SERVICES · KL
Challenge
A consumer lending organisation was preparing to deploy an automated loan eligibility model. Internal confidence in the model's performance was high, but there had been no external assessment of fairness dimensions or alignment with Bank Negara's responsible AI expectations.
Solution
The ethics review examined training data composition across demographic proxies, the decision pathway for borderline applicants, and the model's explainability properties. Stakeholder impact was mapped across applicant segments and the regulatory alignment assessment referenced Bank Negara's TEVR framework.
Results
The scorecard identified two high-priority remediations and four lower-priority observations. The model was adjusted before rollout, and the resulting regulator conversation went smoothly. The organisation now uses the scorecard format for internal reviews of any new AI system before deployment.
Delivered in 4 weeks · MYR 2,200
GET IN TOUCH
33 Jalan Stesen Sentral 5
50470 Kuala Lumpur, Malaysia
Monday–Friday
9:00 AM – 6:00 PM MYT
TRUST CREDENTIALS
MSAI Member Organisation
Malaysian Society of Artificial Intelligence
PDPA 2010 Aligned
Data processing agreements on every engagement
MDEC Digital Nasional Partner
Recognised AI solutions provider, 2023
TEVR Framework Aligned
Bank Negara's responsible AI guidance applied to ethics reviews
YOUR TURN
The best way to understand if Cerebrix is a fit is a brief conversation. There's no pitch involved — just a direct discussion about what you're trying to do.
Start a Conversation