"Token consumption" is the new "Story point".
Companies are literally tracking how many tokens their teams use, like it means something. It doesn't.
Using more tokens doesn't mean better work.
It just means you spent more tokens.
That's like measuring productivity by how many hours someone sat in a chair.
We've seen this before:
- lines of code
- hours in the office
- story points
- number of meetings attended
All garbage metrics that reward looking busy instead of actually doing anything
now it's "ai adoption" measured by token usage.
Same mistake, different dashboard.
I've seen teams burn millions of tokens and produce nothing; literally nothing.
Just people prompting for the sake of prompting because someone told them they need to "use ai more"
The only question that matters is: did anything get better?
Did you ship faster?
Did quality go up?
Did customers notice?
Did engineers stop doing repetitive work they hated?
Did support get less tickets?
Did anything measurable actually change?
If you can't answer that, your ai strategy is just a spending report.
The worst part is when companies incentivize token usage directly. Now people optimize for visible activity instead of output.
They confuse using the tool with getting results.
You're literally training your team to waste money.
Good metrics: time to ship, bugs caught, customer satisfaction, revenue per employee.
Bad metrics: tokens consumed, number of prompts, "AI adoption rate".
Stop measuring whether people touched the tool. measure whether the tool changed the result.
Anything else is noise.