AI is not a strategy.
It is an input. The strategy is still the same one it always was: decide what the business is trying to win at, then choose the capability that gets you there. A company that leads with the tool ends up with a tool.
Insights
What we think, andwhy it matters to you.
These are positions, not articles — each one readable in full, right here. They shape how every engagement gets scoped, so it is worth knowing whether you agree with them before you hire us.
11 — Insights
How we think aboutAI and business.
Positions we hold, plainly stated. They shape how every engagement is scoped — and they are worth disagreeing with before you work with us.
The bottleneck is rarely the model.
In most organisations the constraint is upstream of anything technical — unclear ownership, unagreed definitions, data nobody trusts. The model is the easy part, which is exactly why it gets all the attention.
A pilot that never ships is a cost, not a learning.
Proof of concept is a useful phase and a terrible destination. The work is only real once it is in the flow of work, owned by someone, and measured against the number it was meant to move.
Adoption is a design problem.
Intelligence that arrives in a separate tab gets used in the first week and forgotten in the third. It has to appear where the decision already happens, in the moment it is already being made.
Knowing where AI does not belong is worth paying for.
Some processes are low volume, high judgement, or too consequential to automate. Saying so early is cheaper than discovering it in production, and it is the fastest way to build trust with a sceptical team.
Intelligence compounds; automation does not.
An automated task saves the same hour every time. A system that learns from the business gets better at it. Only one of those is still paying you back in year three.
Long-form writing is in progress. Until it is published, nothing is listed here that you cannot read in full.
Disagree with one\nof these?
Good — that's a more useful first conversation than most. Tell us where we've got it wrong.