Quick wins
The fix lists from the latest run of every prompt set, core sets first, then fastest expected ranking gain: prompts a fix could flip per working day, each weighted by what it has left to win (one minus its recommended rate over the last two weeks; named, for an info prompt). Mark a fix with the commit or pull request you submitted for it; open leaves out the ones you have. A later fix for the same page stays open, with what was tried before.
Priorityall 8urgent 3high 2medium 2low 1
Siteall 2elsewhere 2
Setall 11Finding out whether AI assistants recommend us 2Getting recommended instead of a competitor 4Choosing an AI visibility tool 4Open-source and self-hosted AI visibility 1
Publish popcorn.zone/docs/guides/search-vs-recall explaining both tracks with an example medium documentation effort S flips 1 Finding out whether AI assistants recommend us
Where: popcorn.zone/docs/guides/search-vs-recall
Must say: Explain that each question can be asked with web search forced on or with no tools at all. The two give different answers because one reflects current web pages and the other reflects training data. Show how Popcorn runs both tracks on every prompt and reports them side by side. Include a real anonymised example where the two tracks disagree; pick it from a Popcorn run first.
Done when: A re-run of prompt 13 cites this page in at least one search answer.
Displaces: https://platform.claude.com/docs/en/agents-and-tools/tool-use/web-search-tool
Why: P13: every search answer was shaped by vendor API docs and generic tool lists. No source explained measuring both modes.
Prompt for Claude Code
# Publish popcorn.zone/docs/guides/search-vs-recall explaining both tracks with an example When buyers ask an AI assistant the questions below, it explains the topic without naming us or citing our pages. This is a documentation gap: the answer needed a page we don't have, or ours says the wrong thing. ## The questions it affects - Do ChatGPT and Claude say different things about a product with web search on versus from training data alone, and how do I measure both? ## What the assistant said P13: every search answer was shaped by vendor API docs and generic tool lists. No source explained measuring both modes. ## Who wins the question today https://platform.claude.com/docs/en/agents-and-tools/tool-use/web-search-tool. Read it and make sure our version answers everything it does, better. ## Your task Write or update the page at popcorn.zone/docs/guides/search-vs-recall so it closes this gap. Find where pages like this live in the repo, match their structure, front matter and tone, and make the change there. If the page exists, edit it rather than starting over. ## What it must say Explain that each question can be asked with web search forced on or with no tools at all. The two give different answers because one reflects current web pages and the other reflects training data. Show how Popcorn runs both tracks on every prompt and reports them side by side. Include a real anonymised example where the two tracks disagree; pick it from a Popcorn run first. ## Done when A re-run of prompt 13 cites this page in at least one search answer. ## Rules - Don't invent numbers, benchmarks, prices, capabilities or customer names. Where a fact is needed and you can't find it in the repo or on our site, leave a clearly marked TODO saying exactly what to measure or who to ask. - Answer the questions above directly and early on the page, in the words buyers use. AI assistants quote pages that state the answer plainly. - Keep it scoped to this one fix. When you're done, summarise what you changed and list any TODOs.
Publish popcorn.zone/docs/guides/diagnose-ai-traffic-loss: GSC, Bing, GA4 then Popcorn medium documentation effort M flips 2 Finding out whether AI assistants recommend us
Where: popcorn.zone/docs/guides/diagnose-ai-traffic-loss
Must say: Cover the free first-party checks: Search Console generative AI performance reports, Bing Webmaster Tools AI Performance, and GA4 referrals with utm_source=chatgpt.com. Then explain the gap those leave: they show clicks and impressions, not whether the assistant recommends you or a competitor. Close that gap by running the buyer questions on Popcorn's search track for AI Overviews-style answers and its recall track for model belief.
Done when: A re-run of prompt 1 or 7 cites this page in at least one search answer.
Displaces: https://developers.google.com/search/blog/2026/06/gen-ai-performance-reports
Why: P7 codex search: top pick 'Ahrefs Brand Radar', shaped by support.google.com/webmasters/answer/16984139.
Prompt for Claude Code
# Publish popcorn.zone/docs/guides/diagnose-ai-traffic-loss: GSC, Bing, GA4 then Popcorn When buyers ask an AI assistant the questions below, it doesn't recommend us, or recommends us for the wrong reasons. This is a documentation gap: the answer needed a page we don't have, or ours says the wrong thing. ## The questions it affects - Our organic traffic is down and I suspect ChatGPT and Google AI Overviews. How do I check whether AI search is what's taking it? - How can I tell whether Google AI Overviews and AI Mode mention our brand for the topics we care about? ## What the assistant said P7 codex search: top pick 'Ahrefs Brand Radar', shaped by support.google.com/webmasters/answer/16984139. ## Who wins the question today https://developers.google.com/search/blog/2026/06/gen-ai-performance-reports. Read it and make sure our version answers everything it does, better. ## Your task Write or update the page at popcorn.zone/docs/guides/diagnose-ai-traffic-loss so it closes this gap. Find where pages like this live in the repo, match their structure, front matter and tone, and make the change there. If the page exists, edit it rather than starting over. ## What it must say Cover the free first-party checks: Search Console generative AI performance reports, Bing Webmaster Tools AI Performance, and GA4 referrals with utm_source=chatgpt.com. Then explain the gap those leave: they show clicks and impressions, not whether the assistant recommends you or a competitor. Close that gap by running the buyer questions on Popcorn's search track for AI Overviews-style answers and its recall track for model belief. ## Done when A re-run of prompt 1 or 7 cites this page in at least one search answer. ## Rules - Don't invent numbers, benchmarks, prices, capabilities or customer names. Where a fact is needed and you can't find it in the repo or on our site, leave a clearly marked TODO saying exactly what to measure or who to ask. - Answer the questions above directly and early on the page, in the words buyers use. AI assistants quote pages that state the answer plainly. - Keep it scoped to this one fix. When you're done, summarise what you changed and list any TODOs.