◈ Tessera
Client satisfaction

Client Experiences

What organisations say
after working with us

These are experiences from organisations across Malaysia who have completed engagements with Tessera across our three service areas.

Back to Home

7+

Years of work

80+

Engagements completed

94%

Client satisfaction rate

3

Specialised service areas

What clients have shared

"We had been collecting customer feedback for two years without a consistent way to review it. The NLP analytics engagement gave us a structured taxonomy and a dashboard that now runs as part of our regular reporting. The scope was defined clearly at the start and delivered within the agreed timeframe."

NZ

Nur Zahra

Customer Experience Lead · Petaling Jaya

March 2025

"Our data pipelines were creating delays that affected our weekly reporting cycle. The six-week engagement with Tessera addressed the core issues — schema inconsistencies and an ETL process that had grown unwieldy over time. The monitoring setup was something we hadn't prioritised before, and it has already caught two issues before they became problems."

KP

Krishnan Pillai

Head of Analytics · Shah Alam

February 2025

"The recommendation engine Tessera built for our platform is now handling personalisation for the majority of our product suggestions. The A/B testing framework they set up made it straightforward to evaluate the impact. Communication throughout the engagement was direct and useful — we always knew where things stood."

WL

Wong Ling Fang

E-Commerce Director · Kuala Lumpur

March 2025

"We came to Tessera with a fairly unclear brief — we knew our data was underutilised but weren't sure where to begin. The scoping conversation was genuinely helpful. We ended up starting with the data pipeline service, which gave us a much clearer picture of what our data actually looked like before we committed to anything more complex."

SR

Shazrul Rahimi

Operations Manager · Selangor

January 2025

"The NLP analytics work was used to process support ticket data that had been accumulating for some time. What we found in the outputs — the recurring themes, the sentiment shifts by product category — was useful enough that it changed how we prioritised certain issues. The handover documentation was thorough and our team has since extended parts of the system."

TH

Tan Hui Ling

Product Manager · Penang

February 2025

"We had looked at a few options before settling on Tessera. The fixed-scope model was what drew us in — it made budgeting straightforward and removed some of the uncertainty we had felt about AI projects in the past. The work itself was delivered on schedule and the practitioner we worked with understood our sector."

AM

Azlan Mustafa

IT Director · Kuala Lumpur

March 2025

A closer look at three engagements

Natural Language Analytics · 8 weeks

Processing two years of support ticket backlog

Challenge

A software company had accumulated over 14,000 support tickets across two years. Manual categorisation was inconsistent and the team had no reliable view of which issues appeared most frequently or how sentiment correlated with product version.

What We Did

Designed a taxonomy of 22 issue categories through structured review of a sample set. Built a sentiment model trained on labelled examples and applied it to the full archive. Created a dashboard showing ticket volume, category distribution, and sentiment by product area.

Outcome

The team identified three recurring issue categories that accounted for 41% of all tickets — none of which had been previously tracked as a group. Product prioritisation changed based on the findings. The dashboard continues to update with new tickets on a weekly basis.

Data Pipeline Optimisation · 6 weeks

Reducing a 14-hour reporting lag to under 2 hours

Challenge

A retail organisation's weekly reporting took up to 14 hours to run due to an ETL process that had grown incrementally over five years without structural review. The data team spent significant time each week managing failures and reruns.

What We Did

Mapped the existing pipeline and identified three stages responsible for the majority of processing time. Redesigned the ETL structure around normalised schemas, removed redundant transformation steps, and added monitoring with alerts for processing time thresholds.

Outcome

Weekly report processing dropped from 14 hours to under 2 hours. The team now receives alerts when processing approaches threshold rather than discovering failures after the fact. A subsequent NLP analytics engagement is being scoped using the same data infrastructure.

Recommendation Engine Design · 12 weeks

Adding personalised product suggestions to an e-commerce catalogue

Challenge

An online retailer with approximately 3,000 SKUs was surfacing products manually through editorial curation. The team wanted to introduce automated personalisation but did not have the in-house capability to design or implement a recommendation system.

What We Did

Analysed 18 months of purchase and browsing behaviour. Designed a hybrid recommendation architecture combining collaborative filtering with content-based signals for new users. Built an A/B testing framework to measure the impact of personalised recommendations against the existing editorial approach.

Outcome

The A/B test showed a 23% increase in click-through rate on recommended products over the control group within the first four weeks post-deployment. The system has been operating with the client's team managing updates since handover in early 2025.

Professional standing

MSC Malaysia

Certified Technology Company

PDPA Compliant

Personal Data Protection Act, MY

MDEC Partner

AI Adoption Programme, 2023

4.7 / 5.0

Average client rating

Ready to start a conversation?

If something in these experiences resonates with your situation, we are happy to talk through what might be useful for your organisation.

+60 3-6201 4837
7 Jalan Sultan Ismail, KL