We ask theassistant.Then we publish.
Each report is one real test: we ask an AI assistant to recommend a business in a Sydney suburb, run every prompt several times in a clean session, and publish the prompts, the method and the results. We name the businesses it recommended. We never name the ones it missed.
The first test is being run.
Reports appear here once the test behind them has actually been run. Nothing is published from an estimate, and nothing is listed before it is real. The first is 45 runs across five Sydney suburbs, asking ChatGPT to recommend a dentist.
Run the test on your businessOne profession at a time, in this order. Each is a separate run with its own prompts and its own published method.
Want to know if you get named?
We'll run the same test on your suburb and send you the results - who the assistant names, and where you sit. Free, no call required.