Client Experiences
What clients say about working with us
These are genuine accounts from organisations that engaged Pulsarai for AI consulting and development work across Malaysia.
Back to Home40+
Engagements completed
4.8/5
Average satisfaction score
8
Industries served
6+
Years in practice
Client Reviews
Words from the organisations we've worked with
Hairul Azrin
Operations Manager, Food Processing · Klang
"We brought Pulsarai in to look at our manual quality inspection process on the packaging line. What struck me was how thoroughly they understood our context before proposing anything. The computer vision system they built cut our defect escape rate significantly — and my team actually understands how it works, which matters for day-to-day operations."
January 2025
Nadia Rashid
Head of Analytics, Financial Services · Kuala Lumpur
"We needed to make sense of customer feedback arriving in multiple languages — English, Bahasa Malaysia, and mixed. The NLU platform Pulsarai built handles all of that and classifies sentiment with a level of accuracy we weren't expecting. The handover was thorough and the documentation is genuinely useful."
February 2025
Calvin Siow
CTO, Logistics Technology Firm · Shah Alam
"We had three separate ML models that we couldn't get to work reliably in production. The integration architecture work from Pulsarai untangled the data flow issues and gave us a blueprint we could actually implement in phases. They were honest about what was realistic within our timeline — no overselling."
December 2024
Faridah Mohamad
Director, Healthcare Administration · Petaling Jaya
"Our document digitisation backlog was a real operational drag. The computer vision solution Pulsarai delivered handles our medical forms and referral letters with accuracy we trust. The privacy handling throughout the project was particularly important to us — they understood that from the start."
January 2025
Wei Teng
Product Manager, E-commerce · Subang Jaya
"We needed topic categorisation for tens of thousands of product reviews per month. Pulsarai built a clean NLU pipeline that connects directly to our existing data warehouse. Turnaround time from initial discussion to production was about twelve weeks — they stuck to it."
November 2024
Rajesh Selvaraj
IT Lead, Manufacturing Group · Prai, Penang
"The AI integration architecture work gave our group a clear picture of how to move forward with the three separate AI tools we had purchased but never properly deployed. The phased implementation plan was practical and something our internal IT team could actually execute without external support at each step."
December 2024
Case Studies
Three engagements, in their own words
Computer Vision · Food Processing
Quality control on a high-volume packaging line
Challenge
A food packaging manufacturer in the Klang Valley had a 2.8% defect escape rate on a high-speed line — manual inspection was inconsistent across shifts, and defects were being caught at distribution rather than at source.
Solution
Pulsarai scoped a computer vision system for inline defect detection, built a training dataset from historical defect samples provided by the client, and deployed a model integrated with the existing conveyor control system over a 12-week engagement.
Results
Defect escape rate reduced from 2.8% to 0.4% within the first month of production operation. Inspection throughput increased due to elimination of manual sampling. The system has operated in production for over six months without retraining.
"The system works on both day and night shifts without needing adjustment. That consistency was exactly what we needed." — Operations Manager
NLU Platform · Financial Services
Multilingual customer feedback classification at scale
Challenge
A financial services firm received thousands of unstructured customer feedback submissions monthly across English, Bahasa Malaysia, and mixed-language text. Manual categorisation was creating a three-week backlog in the customer experience team.
Solution
Pulsarai built a multilingual NLU pipeline with custom sentiment and topic classification layers, incorporating the client's internal taxonomy developed through domain expert sessions. Deployed as a REST API feeding directly into the client's CRM.
Results
Processing backlog eliminated. Classification accuracy across languages: 89% on topic, 91% on sentiment. The customer experience team's review workload shifted from raw classification to exception review — reducing processing time by roughly 70%.
AI Integration Architecture · Logistics Technology
Operationalising three undeployed AI models
Challenge
A logistics technology firm had three separate ML models — demand forecasting, routing optimisation, and anomaly detection — that had never successfully reached production due to data pipeline inconsistencies and unclear integration ownership.
Solution
Pulsarai audited the existing infrastructure, identified and resolved data pipeline inconsistencies, and produced an integration architecture blueprint with a phased deployment plan. Each model was given a defined integration specification and monitoring setup.
Results
All three models deployed to production within 14 weeks of engagement start. The phased plan allowed the internal engineering team to implement integration for two of the models independently, using the specifications and patterns provided in the blueprint.
Location
9-B, Jalan PJU 5/1, Dataran Sunway, 47810 Petaling Jaya, Selangor
Your turn
Ready to discuss your AI challenge?
We're straightforward in our initial conversations — no pressure, just an honest assessment of what's feasible for your situation.
Get in Touch