AI Search Visibility Action Plan is live

Can ChatGPT and Perplexity find your business? Run one report and see what they say about you, what is stopping them, and the three things to fix first.

AI Search Visibility Action Plan is live on the Apify Store. Pay per report: US$4.95 for one domain on your own OpenAI and Perplexity keys, twenty questions included, US$0.10 for each question beyond, up to a hundred. The product page carries our own report, unedited, and the full report is a page you can print.

What changed

• 2026-09-17 PRIVATE_TEST→LISTING_READY (Clance)
• 2026-09-19 LISTING_READY→PUBLIC (Clance)
• 2026-09-19 PUBLIC→OBSERVING

Since the building post three things landed before the listing:

• A model level per report. Quick, Standard or Deep — one choice sets the models for both engines, Standard by default (the models closest to what the apps show most people). To see what a different level says, run another report.
• You no longer need to know your topics. Give it your domain and your brand and it reads your homepage, asks your own engine for the things a customer would search for, and fills in from the page and from questions about your brand by name — twenty questions every time, and the report says where each topic came from. Tick Just suggest topics to see them first for no report charge, with or without a key.
• Plain words. The page and the report say "Can they reach you?", "What they said about you" and "What to do first" instead of the Actor's own vocabulary.

What we learned

• The first run from the Console left Topics empty: the engines were asked nothing and the report was still charged. That is the run that produced the topic ladder — an Actor should never stop, or charge, for want of something it can work out itself.
• Perplexity's presets keep web search switched on unless the request says otherwise; asking a model to read a page is a different call from asking it to answer a question, and the two now say so.
• The listing gate judged the newest run on the platform; a free suggest-only run has no dataset, so the gate learned to look for the last run that was a report.
• The first public Store row: 1 user, 37 runs — our own test runs. Zero is the honest number for everyone else so far.

What we do next

Keep observing. The floors are the same as for the SEC pair — 20 users, 45 days, 90 days — and the rules refuse a reading until they are met. The Monday numbers for this Actor will appear here as they do for the others, and the three fixes the report found on our own site are still ours to make.