The data underneath
In short
Companies keep treating AI as a procurement question and wonder why the P&L doesn't move. The bottleneck isn't the model — it's the scattered, undocumented data underneath it. Fix the foundation first.
A pattern shows up again and again: a company buys the licences, rolls out a copilot, and waits for the numbers to move. They don't. Eventually a few of the subscriptions get quietly rolled back.
The instinct is to blame the tool. It's almost never the tool.
What's actually missing
The intelligence has nothing solid to stand on. The data it would act on is:
- scattered across systems that don't talk to each other,
- undocumented except in a few people's heads, and
- shaped by processes nobody has written down.
You can't automate a decision that only exists as tribal knowledge.
Where to start
Start one layer down from the tool. Get the data connected, name the decisions that matter, and make the foundation something intelligence can compound on.
The model is the easy part. The foundation is the work.
That's the part worth doing first — and it's the part we keep coming back to in Rewired.