— Our Story
How Tepi came to be
Tepi Technologies grew from a frustration shared by several colleagues who had watched AI projects arrive at organisations, generate excitement, and then quietly recede — leaving little behind. The workflows built without adequate documentation were difficult to maintain. The staff trained in one-day workshops forgot most of what they had learned. The value promised did not persist.
We started Tepi in Singapore because the city's organisations — spanning finance, professional services, logistics, and the public sector — are genuinely curious about AI but often underserved by engagements that move faster than understanding can form. We wanted to work differently: more slowly, more carefully, and with a commitment to leaving behind written records that would remain useful after we had gone.
The name Tepi is drawn from the Malay word for edge or margin — a nod to the design language we use and to our conviction that the most important things in any document are often found in the annotations, the caveats, the notes at the side of the page.
Our Mission
"To help Singapore organisations integrate AI in ways that are understood by the people who use them, documented well enough to survive staff turnover, and honest about their limitations."
Founded
2022, Singapore
Headquarters
71 Robinson Road, #14-01, Singapore 068895
Practice Area
AI Integration & Workflow Documentation
— The People
Who works at Tepi
Wei Liang Ong
Principal & Co-founder
Wei Liang leads client engagements and is responsible for the written documentation that accompanies each build. He spent eight years in process consulting before co-founding Tepi.
Siti Rahimah Binte Aziz
Practice Lead, Workflow Design
Siti designs and builds the AI workflows, working closely with client staff throughout the build phase and the four-week trial. Her background is in operations management and NLP.
David Muthusamy
Practice Lead, Client Relations
David handles initial conversations with prospective clients, scoping, and ongoing communication during engagements. He has a background in management consulting and technical writing.
— How We Work
Standards we hold ourselves to
These are not policies for display. They reflect how we have chosen to practise, and we are happy to be held to them.
Written deliverables, not verbal summaries
Every engagement produces a written document. We do not deliver findings verbally and consider the engagement closed. Our written outputs are edited for clarity and intended to be useful to non-technical readers.
Data handling with a written agreement
Before any work begins, we agree in writing on what data we will see, how it will be handled, and the conditions under which it will be destroyed. We do not retain client data after an engagement closes.
Honest about what AI cannot do
Each workflow document includes a section describing conditions under which the workflow should not be relied upon. We include this because we think it is necessary, not as a liability disclaimer.
Knowledge transfer as a first-order goal
We measure the success of an engagement partly by whether your team can describe, explain, and maintain what we built together, without our ongoing involvement.
No upselling during engagements
We propose the scope of an engagement at the start and do not expand it mid-engagement without your explicit request. If we identify additional areas of interest, we note them in writing and leave the decision entirely with you.
Fortnightly written notes during programmes
For multi-month engagements, we prepare written notes after each fortnightly session. These notes form a record of decisions made and a resource for staff who join the team mid-engagement.
— Our Practice
AI integration in Singapore — what we have learned
Singapore's professional services sector has been an early and serious adopter of AI tooling. Across legal practices, accounting firms, logistics operators, and financial institutions, there is genuine appetite for workflows that reduce the burden of repetitive, text-heavy tasks. What we observe repeatedly, however, is that adoption without documentation produces fragile results: workflows that depend on one knowledgeable person, that cannot be explained to an auditor, or that are quietly abandoned when something goes wrong.
Our practice addresses this directly. We bring to each engagement a discipline of writing — of producing clear, plain-language records of what is being built, why each choice was made, and what the resulting workflow cannot be trusted to do. This documentation is not a byproduct of our work. It is the work, alongside the workflow itself.
We work with organisations of varying sizes, from small professional practices exploring their first AI workflow to larger enterprises managing a portfolio of connected processes. In each case, the approach is the same: observe first, recommend carefully, build alongside the team, and document everything well enough that our presence is eventually unnecessary.
— Work With Us
Let's begin with a conversation
We are happy to spend thirty minutes learning about your organisation before proposing anything. No fee, no obligation.
Contact Tepi Technologies