Insights

How we think about it

Short written pieces on specific problems — no frameworks, no jargon, no claims we cannot back up in a room. Written from inside engagements, not from a desk.

Latest piece

Forward Look6 min read

What Good AI Governance Looks Like for a 30-Person Company

You don't need a committee. You need three clear decisions documented where everyone can find them.

The assumption that AI governance is an enterprise problem is wrong. Smaller companies face the same risks — data leakage, hallucinated outputs, compliance violations, vendor lock-in — but without the infrastructure to absorb them. A 30-person company can't hide a mistake in a compliance department. The founder's phone rings directly.

2 April 2026Read the piece
What Good AI Governance Looks Like for a 30-Person Company

All pieces

Why AI Projects Fail in Singapore
Market Reality5 min read

Why AI Projects Fail in Singapore

It's not the technology. Most AI projects fail before a single tool is deployed.

The problem shows up three meetings in. Everyone agrees on the goal. The tool gets selected. The budget gets approved. Then, quietly, nothing happens. Not loudl…

November 2025Read
The 6-Month Lifecycle of an LLM Wrapper
Field Note6 min read

The 6-Month Lifecycle of an LLM Wrapper

They demo well. They deploy quickly. And then, around month three, something quietly stops working.

The pattern is consistent enough now that we can describe it before it happens. A company selects an LLM-based tool — chatbot, document processor, internal assi…

December 2025Read
Three Signs Your Business Isn't Ready for Automation
Diagnostic5 min read

Three Signs Your Business Isn't Ready for Automation

Before any tool is selected, three operational conditions predict whether automation will help or hurt.

Sign one: your team can't describe the same process the same way twice. If two people doing the same job give different answers about how that job works, you do…

January 2026Read
Why Agencies Cannot Automate to a Deadline
Field Note5 min read

Why Agencies Cannot Automate to a Deadline

They promise clients faster turnaround. Then they discover their own process cannot survive the rigidity of automation.

The conversation usually starts around a growth target. The agency has scaled headcount, won larger clients, or taken on retainer work that demands predictabili…

February 2026Read
Why Hiring AI Talent in Singapore is Harder Than Buying the Tools
Market Reality6 min read

Why Hiring AI Talent in Singapore is Harder Than Buying the Tools

The tools are on the shelf. The people who can make them work are not.

The tool procurement cycle in Singapore is fast. A decision-maker sees a demo, runs it past finance, and the subscription is live within a fortnight. The people…

March 2026Read
The Vendor Proposal That Looks Like a Solution But Is Not
Diagnostic5 min read

The Vendor Proposal That Looks Like a Solution But Is Not

The proposal is impressive. The scope is clear. The timeline is specific. None of that tells you whether it will actually work.

Vendor proposals for AI implementations follow a predictable structure. The problem statement is articulate. The proposed solution maps neatly to it. The timeli…

March 2026Read
The Hidden Cost of Tech Stack Sprawl
Field Note5 min read

The Hidden Cost of Tech Stack Sprawl

Every new tool promises efficiency. What accumulates silently is coordination debt.

The average company we assess has between twelve and eighteen active SaaS tools. Often, nobody can produce a complete list without checking billing records. The…

April 2026Read

Ready to assess

If what you have read resonates, the next step is simple.

No commitment. A structured conversation to determine whether your conditions are right for what you are trying to do.

Request a Readiness Conversation