Testing an AI teacher and task assistant before you commit

Value at this stage is not a feature count. It is whether one described task comes back as a usable brief with options and live prices, and whether the session teaches you something you can repeat without the tool. A startup weighing an alternative should judge the first few runs on that alone, before any tier question arises.

Each session is recorded with replay, which matters more than it sounds. When you are testing outreach on a small sample, the replay lets a co-founder see exactly what was asked and what came back, so the decision rests on the transcript rather than on your summary of it. Finished runs are also saved as lessons.

Four modes are switchable mid-session, so you can move from being taught to watching the task done, or to doing it alongside the assistant, without starting over. That flexibility is the practical argument for the product here: you learn the method on a real task instead of a training exercise you then have to translate.

Two lifetime tiers are published, Learner and Certified, and the line between them is the decision you actually face. Prices are shown separately in the publisher's own table, so compare the tiers on what each one covers rather than on a figure repeated here. Read the tier descriptions slowly and note which one matches how you plan to use it.

The shape of the payment is the point for a startup. A lifetime tier is a single commitment rather than a renewal that keeps running while your test is still inconclusive. That suits an early team whose spending needs to map to a decision, not to a calendar. Confirm what each tier includes before you pick one.

Do not treat the higher tier as the safe default. If your test is three conversations and a short list of prospects, the question is which tier supports that work and gives you replay and saved lessons you will actually revisit. Pick the tier that matches the test you can run now, not the one that assumes success.

Check the account boundary first. You log in to your own accounts, so the assistant works through access you already hold rather than through credentials it stores for you. Make sure the accounts you connect are ones you are willing to use for a test, and that a co-founder can revoke access if the run does not go as planned.

Check the payment boundary next. The product never enters card or identity details, which keeps those steps in your hands throughout. For a startup that has not yet settled who holds the company card, that division is worth confirming with whoever controls spending before the first session begins, not after.

Check your expectations about completion. Completion is not guaranteed, so some sessions will teach the method and leave the final version to you. Plan the test so a partial run still tells you something useful, and keep a second attempt in reserve. Judge the tool on the quality of what it teaches, not only on tasks finishing.

Decide on the upgrade after you have run a real task, not before. Take one outreach question you genuinely need answered, describe it, and work through it in a mode that suits you. Then replay the session and ask whether the brief, the options and the live prices gave you something you could act on without further help.

If the answer is yes, the upgrade question becomes simple: does the higher tier cover the way you now know you will use the product? Re-read both tier descriptions with your own session in mind. A tier that adds coverage you will not touch is not an upgrade for your situation, whatever the label suggests.

If the answer is no, do not upgrade on the strength of one poor run. Try a different mode mid-session, or describe a narrower task, and see whether the teaching improves. The upgrade decision should follow evidence from your own sessions, and the replay and saved lessons give you that evidence in a form a co-founder can review.