A typical first month
Most people use Proofoak in short, focused sessions of 30–90 minutes, a few times a week. Here's what a first month usually looks like.
| Week | Sessions | What you finish |
|---|---|---|
| 1 | Starting point, idea library or an Idea Check | A chosen model or a checked idea, with its risks and a first test |
| 2 | Kit, offer, practice | A versioned one-page offer and the contact messages you'll use |
| 3 | Test plan, first contacts | An approved test with a locked rule; messages sent from your own account |
| 4 | Evidence, decision | A decision against your rule and the next version of your offer |
You can move faster or slower. The workspace always shows the next task and why it's next, so returning after a week doesn't mean starting over.
What you do, and what Proofoak does
| Proofoak | You |
|---|---|
| Brings you researched models and checks your own ideas with live, sourced research | Give accurate information and choose |
| Pre-fills the offer, test plan and contact messages from a kit or an Idea Check | Learn, correct and approve |
| Shows where your first buyers are and which channels are permitted | Decide who to contact |
| Locks your decision rule when the test starts | Send messages yourself through permitted channels |
| Organises evidence per person and compares it with your rule | Hold the conversations and make the decision |
| Never sends, publishes or spends for you | Deliver the service to your own client |
Free, Build and Launch in the same project
On Free, you can complete your starting point, read every dossier summary, open one ready-to-run kit a month, run one quick Idea Check a month and run a first test by hand with up to 10 evidence records. Build opens every kit and five full Idea Checks a month with live web research, and Launch adds early access to new dossiers, market adaptation reports and three projects at once. Upgrading never restarts your work, and every check shows how many you have left before you run it.
Safeguards built into every step
- Versions, not overwrites. Offers and profiles are saved as versions; a test always knows which offer it used.
- Rules before results. The decision rule locks when the test starts, so it can't be rewritten to force a positive outcome.
- Labels on every claim. Facts carry a source; estimates and hypotheses are marked as such.
- Honest outcomes. "Not run" and "inconclusive" are real results, not failures to hide.
- Your data stays yours. Export everything at any time. Your ideas are sent to our AI provider only to write your report and are not used to train models.