Peter Holford’s AI studio — apps, decision tools, and working notes on deciding what to build.
Two decades shipping at the front of each wave — ecommerce, mobile payments, open banking, big data — at Worldpay, Experian, HSBC and NatWest, then venture building at Prehype. Each wave changed what you could build. This one changes how you decide. This studio is where I work that out in the open: build, test, publish.
Two products, both now in beta test and open to try. Beta means the feature set is still moving — not that the build is rough: these are production-grade and properly secured, and your data is handled the same way it will be at launch. Both went through the process below before I committed to building either.
Founder discipline is the one thing AI has not made cheaper. It matters more now that anything can be made to look finished in an afternoon.
A good idea does not mean a good product.
A good idea does not mean a good business.
Those are two different failures, and most ideas die of one or the other. Venture Builder is how you find out which one you’re facing, before the building starts.
Where an idea gets assessed, tested and de-risked: assumptions named, experiments run against them, and decisions logged with the reasoning at the time. It’s how the apps above get built — Aida and Sorted Careers both run on it. It’s also a standalone product, available to use on your own ventures.
Open the workspace →Each tool answers one question the process asks. They work inside a venture, or on their own. Free during beta.
Ideas worked through far enough to be worth arguing with, published to invite collaboration rather than to settle anything. Split by how finished the thinking is — and if you think something here is wrong, I'd like to hear it.