Enterprise AI

The interesting questions are about governance, evidence and decision quality — not capability.

The honest position

Capability is no longer the constraint for most organisations. What constrains them is the same thing that constrained the previous three waves of enterprise technology: unclear ownership, poor data discipline, and no agreed standard for what counts as a good enough answer.

The question I ask first

Which decision is this intended to improve, who currently makes it, and how would we know afterwards whether it improved? Where those cannot be answered, the initiative is a pilot in search of a sponsor, and it will be quietly discontinued in a year.

Governance is the real work

An organisation adopting AI inside its operations is extending decision-making authority to a system whose reasoning it cannot fully inspect. That demands the ordinary disciplines — traceability, accountability, review, the ability to say why a decision was made — applied more strictly, not less, because the failure mode is confident and plausible rather than obviously broken.

What I am deliberately not claiming

This page is short on purpose. I would rather state a defensible position than assemble a capability narrative around work that has not yet been done at the scale such a claim would imply.

The argument behind this, and where it is still being worked out, sits in thinking.