Anvil
Six weeks to an AI strategy you can defend.
What is Anvil?
A lot of AI efforts start with a tool. Someone buys a license, a pilot runs on enthusiasm, and six months later leadership is still asking what the AI strategy is. Anvil starts with the problems instead. Your team picks real problems, tests a few ways to solve each one, and builds only what the evidence supports. The strategy we hand you is made of that work.
What you get
A strategy you can defend
We write it from six weeks of real work inside your organization. It covers what was tested, what the evidence showed, and what to do next, and we present it to leadership in a one-hour readout.
Working prototypes
Each participant builds one, aimed at a real problem and usable by someone besides the person who built it. Most are things like a Claude Project or a small template system. No coding required.
A repeatable method
Your people learn to find the right problem, run small experiments, and decide on evidence. They keep using it with AI after we leave.
How it works
Each person brings one problem. In weeks one and two they find the real problem under the symptom and check it with the person who has it. In weeks three and four they test three ways to solve it, and nobody builds until the evidence points somewhere. In weeks five and six they build the smallest real version and demo it. The thinking behind the first two weeks is in Fall in Love with the Problem.
TWO FORMATS
Six weeks or five days
Anvil runs six weeks at two hours a week, plus office hours. Anvil Intensive covers the same program in five full days. It's built for two to eight people from one organization. You'll need a computer, a Claude account (Pro is enough to start), and a problem worth six weeks.