Operator work for AI that has to pay off
The straight read on what actually works.
Most companies have already bought the AI. Few are getting anything back. We find where the value is stuck and build the fix. Then we map where it breaks next, so you plan around the limits instead of hitting them in production. Everyone else sells the edge. We show you where it breaks.
Latest writing
All writing →- aiJul 2026
Developers Using AI Were 19% Slower. They Reported Being 20% Faster.
Adoption is a measurement claim, and fewer than one in five companies own the instrument that would let them make it. The one controlled trial that put a stopwatch next to the feeling found the two disagreed by 39 points, which is a problem for every self-reported AI productivity number in circulation.
- aiJul 2026
A Pipeline With No Agents Beat Every Agent Framework. It Cost 70 Cents a Task.
Teams are buying autonomy they do not need and skipping the structure that demonstrably works. Anthropic's own numbers put multi-agent systems at 15x the tokens of a chat, with token spend alone explaining 80% of the performance gain. Meanwhile the one controlled test of whether scaffolding becomes unnecessary as models improve rejected its own hypothesis.
- aiJul 2026
88% of Companies Use AI. About 5% Get Anything Out of It.
Adoption is nearly universal and value capture is rare, and the gap is not the model. Across McKinsey, BCG, Stanford, and MIT, the one thing that separates the companies getting returns is that they redesigned the work, not that they bought a better model. Here is what the controlled evidence actually supports, and where it runs out.
How this works
The operator work. We audit where your AI investment is stalling, build the systems and automation that unstick it, and scope the engagement to the problem rather than a fixed package.
The writing is the proof. Every piece states the question, what the evidence holds, and where it runs out. Method plus a stated failure boundary is the line between analysis and marketing. It is how we think, in public.
Commissioned research. When a question needs original work, we run it under the same method, on your problem, with the failure boundary stated.