Describe your campaign and this assumes it already failed, then reasons backward to the most likely reasons: hidden assumptions, ranked risks, the early signals to watch, and what to fix first. Risk analysis, never a prediction.
Surfaces risks & signals — never invents benchmarks or predicts outcomes.
Every account gets one free run, then it's 1 credit each (free monthly credits cover normal use).
The opposite of a post-mortem: instead of analysing a failure after it happens, you assume the campaign has already failed and work backward to why — so you can fix those things before spending the money. It's one of the highest-ROI 10 minutes in marketing.
No — and it won't pretend to. It surfaces risks and the early signals to watch, never a forecast or invented benchmark. Thresholds are left as ranges or [blanks] for you to set from your own targets (e.g. your max CPA). It's risk analysis to act on, not a prediction.
It doesn't write your campaign — it stress-tests the one you have. It's adversarial by design: its job is to find what could go wrong so you fix it before launch.