Happy to sign your NDA before anything is shared. Prototypes run on a private, login-protected environment, and I work with anonymised or synthetic data by default unless you’d rather supply a sanitised extract. If your team has its own stack and approval process, I’ll work inside yours instead of asking you to adopt mine.

By whether you can make a decision at the end of it. Every prototype ships with comments and reviews attached, so you finish with a prioritised list of what real users tripped over and what your own team flagged. If the prototype proves the idea is wrong, that still counts as a result. Finding that out in two weeks is the entire point.

Revisions aren’t counted, they’re the process. Each week runs the same loop: new build goes out Monday, comments and review findings come back by Wednesday, revised build follows. That’s two full loops on a two week sprint and four on a four week one.

You get the working prototype, not a description of one. That means the deployed build, the source, the component structure and the tokens behind it, plus a walkthrough video covering the decisions and edge cases the screens don’t explain on their own. Your engineers can read it, lift what’s useful and bin the rest. Nothing is locked to my tooling or my accounts.