What Clients Say About Working With Prismata
Direct feedback from organisations across Singapore's financial, healthcare, and technology sectors that have engaged Prismata for AI engineering work.
Back to HomeCompleted engagements
Average satisfaction score
Years operating in Singapore
Industry sectors served
In Their Own Words
"The data pipeline Prismata designed addressed problems we'd been working around for over a year. The scoping process was thorough — they asked questions our previous vendors never thought to raise. The documentation they handed over was genuinely useful, not the usual placeholder content."
Timothy Wong
Head of Data, Asset Management · Singapore
January 2026
"Our compliance team was spending too much time on manual document review. The monitoring system Prismata built has reduced that workload significantly while keeping our officers in control of every decision. The audit trail implementation specifically was something we'd been wanting for two years."
Sunita Bose
Chief Compliance Officer · Insurance · Singapore
February 2026
"We'd tried to build our AI application internally before engaging Prismata. The difference in the final result was considerable. What stood out most was the review gate process — being able to adjust direction at defined points without disrupting the whole engagement was important for a project of this complexity."
Marcus Loh
CTO, Fintech Scale-Up · Singapore
January 2026
"The pipeline architecture engagement gave our engineering team a clear implementation roadmap. I appreciated that Prismata was direct about what our current setup could and couldn't support — no overclaiming. Our team has been working from their blueprints for three months now without needing to go back to them for clarification."
Priya Nambiar
VP Engineering, Healthcare Group · Singapore
December 2025
"What I valued most was that we were communicating directly with the people building the system, not a project manager passing messages between teams. It made technical decisions faster and the entire engagement felt more like an internal collaboration than an external vendor relationship."
Aaron Chia
Director of Operations, Private Bank · Singapore
February 2026
"The twelve-week post-launch period has been more valuable than we anticipated. Having the same engineers available to iterate on the system based on actual user behaviour — rather than starting a new engagement with someone unfamiliar with what was built — made a real difference to the quality of the final product."
Yong Ting
Product Lead, Enterprise Technology · Singapore
January 2026
Selected Project Outcomes
A closer look at three engagements and the specific challenges each organisation brought to Prismata.
ML Pipeline Redesign for a Singapore Fund Manager
Challenge
The quantitative team's data pipelines had evolved organically over four years. Training data and inference data were being prepared by different processes, creating model performance inconsistencies that were difficult to diagnose.
Solution
Prismata redesigned the pipeline topology to unify training and inference paths, with explicit feature versioning and data quality checkpoints at each stage. The design was handed over with full implementation documentation.
Result
Training-serving skew eliminated within two weeks of the client team implementing the new design. Model performance variance reduced by approximately 40% over the following quarter. Engineering team operating independently within one month.
"The scoping process alone surfaced issues we hadn't realised were connected." — VP Quantitative Research
Compliance Monitoring for a Mid-Size Insurer
Challenge
A compliance team of four was manually reviewing a growing volume of communications and transaction records against MAS conduct requirements. Coverage was incomplete and the team was spending 70% of capacity on routine document review.
Solution
Prismata designed and built an AI-assisted monitoring system that processes communications and flags items requiring human review, with a complete audit trail. The system integrates with the team's existing case management workflow.
Result
Routine document review time reduced from 70% to approximately 30% of team capacity. Coverage of communications increased to 100% from a previous 40%. The team now focuses human attention on the items the system surfaces rather than trying to review everything manually.
End-to-End AI Application for a Lending Platform
Challenge
A consumer lending platform needed an AI-powered credit assessment tool that could handle their specific customer profile while meeting MAS credit guidelines and producing explainable outputs for their compliance function.
Solution
Full-stack development engagement covering model development with explainability requirements, a reviewer interface for credit officers, backend integration with existing systems, and a twelve-week post-launch improvement period.
Result
Application deployed within fourteen weeks, with post-launch improvements to the reviewer interface completed during the support period. Credit assessment consistency improved substantially, and the explainability outputs satisfied compliance review requirements without modification.
Reach Prismata Directly
The most direct way to understand whether an engagement makes sense for your situation is a short conversation with one of our engineers.
Send a MessageAdd Your Organisation to the List
If you have an AI engineering challenge that requires careful, structured work — we'd like to understand it.
Start a Conversation