From the assessments
What the bill does not say
Your model invoice knows tokens and models. It knows nothing about workflows, teams or outcomes. Which is why it cannot answer the one question the board asks.
We write down what we see.
An insight is either the result of an engagement or the result of our assessments without one. We learn from both.
Why this is here
Anyone who argues that companies cannot interpret their own AI costs has to publish their own methods. Otherwise it is the same black box in a different colour.
There is no metric here whose origin we cannot state. Where a derivation or the evidence is missing, no number is given.
From practice
The first meter hung on our own AI
Tokenometrics did not come out of a workshop. It exists because we could not explain our own invoice, and then found the same gap at every client we looked at.
From the assessments
AI and vibe coding: what of it survives production
Five claims about AI in software development, held against what we actually use. Including where the answer is uncomfortable.
Questions about the Insights?
Disagreement is welcome, especially about the method. I'll get back to you within 24 hours.
