INSIGHTS

No theory. Just what works.

What we've learned building AI tools, rescuing stalled projects, and getting data to behave — from the team doing the work.

The early signs a software project is in trouble

20 Nov 2025 · Gerald Yeo

The early signs a software project is in trouble

Failing projects send signals long before the deadline is missed. The early warning signs, how to read them, and why the symptom is rarely the real problem.

DeliveryEngineering
Read more
Running AI agents in production

28 Oct 2025 · Yaohong Ch'ng

What it takes to run AI agents in production

A demo agent only has to look right once. A production agent has to be safe every time it acts. The guardrails that close the gap: bounded scope, least privilege, checkpoints, and observability.

AIOperations
Read more
Single source of truth

5 Oct 2025 · Yaohong Ch'ng

Single source of truth: what it really takes

A single source of truth isn't a database you buy. It's the discipline of writing down what's true and keeping it that way, and AI is making it unavoidable.

Data
Read more
The real cost of skipping tests

12 Sep 2025 · Gerald Yeo

The real cost of skipping tests

'We'll add tests later' trades a small cost now for a larger one later. What that deferred cost actually is, and how to spend testing effort where it pays rather than chasing a coverage number.

EngineeringQuality
Read more
Getting your data AI-ready

20 Aug 2025 · Yaohong Ch'ng

Getting your data AI-ready

Most AI ambitions run aground on the data, not the model. What 'AI-ready' really means: structure, movement, and a version of the truth you can trust.

DataAI
Read more
Habits that keep software shippable for years

28 Jul 2025 · Gerald Yeo

The boring habits that keep software shippable for years

Durable systems come from boring decisions made consistently: small changes, a real test safety net, boring technology, and code written to be read. The habits that compound.

Engineering
Read more
Why AI projects fail before production

5 Jul 2025 · Yaohong Ch'ng

Why AI projects fail before they reach production

The demo is the easy part. Most AI initiatives stall in the move to production, on reliability, engineering, and oversight, not on the model itself.

AIStrategy
Read more
Where to start with AI

12 Jun 2025 · Yaohong Ch'ng

How to use AI in your business: start where being wrong is cheap

The most visible AI project is the worst place to begin. Start with a task where being wrong is cheap and the result is verifiable, then escalate from there.

AIStrategy
Read more

Let's figure it out together.

No pitch deck, no pressure. We'll talk through your situation, share an honest perspective, and tell you if we're not the right fit.