18 April 2026
The tooling works. The operating model has not caught up.
The tool produces an answer in four seconds. The organisation still takes four weeks to act on it. The bottleneck has not moved. It has just got more visible.
I keep being asked to diagnose stalled AI initiatives. The conversation almost always starts at the tool. The tool is fine. The tool is doing what the brief asked. The tool is not the problem.
The problem is sitting one layer up: the operating model. Decision rights that were drawn for a world where decisions took two weeks. Review cadences that were set when reports came out monthly. Job descriptions that name the activity, not the outcome. Performance frameworks that reward what was easy to measure five years ago.
You can lay the best AI tooling over that operating model and the result is friction. Not transformation. The tool produces an answer in four seconds. The organisation still takes four weeks to act on it. The bottleneck has not moved, it has just got more visible.
The work is in the operating model. It always was. AI is the X-ray that shows where it was already broken.
More from the Friday Frame archive.
- 9 October 2026
What’s changed since we approved it?
Does the approval still fit the work?
- 2 October 2026
What changes after the training?
Training becomes useful when people have the workflows, incentives and support to put what they learn into practice.
- 25 September 2026
A shared platform still needs local judgement.
Scaling across markets means deciding what stays consistent, what local teams can adapt and who has authority to act.
Get a heads up when there is new writing here.
I publish irregularly. The Friday Frame ships most weeks. The longer essays land when they are ready. Drop your email and you will get a note when something new goes up, or when an existing post gets a meaningful rewrite. Nothing else.