// Client Feedback

What clients say about the work

Experiences from organisations across financial services, healthcare, and logistics that engaged Dataveil on privacy, performance, and intelligence challenges.

Back to Home

47+

Engagements completed

4.9

Avg. satisfaction score

Industry sectors served

98%

Delivered within scope

// Reviews

Client testimonials

"The privacy consultation gave our architecture team a concrete framework they could actually implement. The document wasn't a generic PDPA checklist — it addressed our specific challenge of embedding differential privacy into our customer scoring model. That level of detail was what we needed."

Tan Jiaming

Head of Data · Fintech Platform

February 2026

"Our inference pipeline was running at three times the latency we needed under peak load. The changes brought it down significantly. The benchmarking report was detailed enough to satisfy our CTO and our risk team. The initial scoping conversation could have been slightly more structured, but the technical outcome was solid."

Priya Krishnaswami

ML Engineering Lead · Logistics

January 2026

"We had good data but only two people in the organisation could use it. After the BI Suite engagement, our clinical operations director can ask questions about patient flow data directly and get coherent answers. The training sessions were well paced and genuinely addressed what our less technical staff needed."

Marcus Wong

CTO · Healthcare Administration

February 2026

"What set this engagement apart was honesty during scoping. Dataveil told us upfront that one of the three things we asked about wasn't in their remit, and pointed us to someone better positioned for it. That directness made us much more confident about the parts they did take on."

Siti Rahayu

VP Technology · Insurance Services

January 2026

"We needed documentation showing our AI model met PDPA requirements before taking it to production. The consultation gave us exactly that — written in language both our engineers and compliance function could work from. The vendor evaluation criteria section was particularly valuable when we started procurement."

Aditya Kumar

Head of AI Products · Payments

February 2026

"Latency came down meaningfully and the infrastructure cost reduction was visible within the first month. The documentation handed over is something our DevOps team still references regularly. If I had to name one thing to improve, it would be slightly more frequent updates during the middle weeks of the engagement."

Li Xuefeng

Platform Engineering · E-commerce

January 2026

// Case Studies

Engagement stories in depth

// Challenge

Financial services firm building credit model with personal data

The organisation needed to demonstrate PDPA and GDPR readiness before a European partnership could proceed. Their data science team had technical capability but no established privacy-by-design process.

// Solution

Privacy-Preserving AI Consultation

A 2.5-week engagement reviewed the model architecture, data pipeline, and retention practices. Differential privacy techniques were assessed for the specific credit scoring context and a structured technical guideline was produced alongside vendor evaluation criteria.

// Outcome

Compliance documented, partnership progressed

The guideline satisfied the European partner's data protection review. The model moved to production within six weeks of the engagement closing. The compliance team now uses the document as a template for subsequent AI projects.

2.5 weeks · SGD 420

// Challenge

Logistics operator with demand forecasting model underperforming at scale

The model worked well during pilot but degraded significantly under live traffic during peak fulfilment periods. Infrastructure costs were also growing faster than query volume, suggesting deployment inefficiency.

// Solution

Real-Time AI Inference Optimisation

Profiling identified the bottleneck was in the pre-processing pipeline rather than the model. Batching strategy was restructured, a caching layer introduced for repeated queries, and the model quantised without accuracy loss. Changes were staged under simulated peak conditions before deployment.

// Outcome

62% latency reduction, costs stabilised

Latency under peak conditions dropped from 380ms to 145ms. Infrastructure costs as a percentage of query volume stabilised within the first month post-deployment. The benchmarking report was presented at a board technology committee.

5 weeks · SGD 1,180

// Challenge

Healthcare group with rich data only two people could use

Three years of operational and clinical data sat in a well-maintained warehouse, but the BI platform required SQL proficiency. Clinical directors relied on scheduled reports that were often a week old by decision time.

// Solution

AI-Integrated Business Intelligence Suite

A natural language querying layer was embedded into their existing platform. Anomaly detection was configured for three key operational data streams. Predictive widgets were added to existing dashboards. Role-based access ensured clinical staff could see patient flow data without touching financial records.

// Outcome

Decision cycle shortened from days to same-day

Twelve staff who previously relied on weekly scheduled reports can now query data independently. Time from data event to actionable insight shortened from approximately five days to same-day in most cases. Three anomalies were detected and acted on within the first month of going live.

10 weeks · SGD 2,280

// Credentials