Win recommendations:
Prompt intent, differentiation and citations
Winning means being on the shortlist for the buying prompts in your category. Map the topic, research prompts by intent, close citation gaps with owned pages and earned mentions, and use ChatGPT ads where organic answers are out of reach.
Winning recommendations:
Three moves that put you on the shortlist
How do you go from mentioned to recommended?
Engines recommend the brands that fit the stated need and that the sources they read agree on. So target buying prompts, make your fit obvious, and get onto the sources engines cite for those prompts.
- Prompt intent: which buying questions you target.
- Differentiation: who you are best for, in plain words.
- Citations: presence on the sources engines read for those prompts.
Good sign: your share of voice rises on unbranded "best X" prompts.
Warning sign: you are named only when the prompt already includes your brand.
pillars in this course, from research to attribution
This courselessons and research pages, each answering one cluster
This coursemonthly searches for "chatgpt ads", the biggest keyword in the map
DataForSEO, Nov 2026of ChatGPT prompts matched no traditional keyword (vendor claim)
Semrush, Apr 7, 2026Lessons in Win recommendations.
- Lesson 4.1→
Build a topic map
The topic map for the AI search optimization course: 6 pillars, 33 lessons and the keyword data behind them, from chatgpt ads to query fan out. - Lesson 4.2→
Research buyer prompts
Prompt research turns a keyword into the full questions buyers ask ChatGPT and Perplexity. See how to pick, tag and track them, with an interactive demo. - Lesson 4.3→
Win best X prompts
Win best X, vs and alternatives prompts in ChatGPT and Perplexity: the pages to build, the sources to earn and how AI engines pick products. - Lesson 4.4→
Find citation gaps
Find AI citation gaps: compare the sources engines cite for your prompts, see why competitors appear in AI answers, and close each gap. - Lesson 4.5→
Close gaps with owned pages
Use prompt runs to find citation gaps, then publish the owned pages that close them: comparisons, guides and answers built for the prompts you lose. - Lesson 4.6→
Fix crawled, uncited pages
Use AI crawler server logs and a content audit to find pages AI reads but never cites, then fix them first. They are the cheapest AI search wins you have. - Lesson 4.7→
Run ChatGPT ads with MCP
Use the ThriveStack ChatGPT Ads MCP to see who advertises on your prompts, which prompts you lose, what rival ads say, and draft ads to publish in ChatGPT. - Research 4.8→
Build an AI prompt panel
Our AI prompt panel research: build a panel from 12 brand signals, weight it to buying intent and run each prompt 7 times a day per engine. - Research 4.9→
Get cited by Gemini
Gemini cites pages it finds through Google Search, and so do AI Mode and AI Overviews, but each picks different sources. See what works on each.
See this hub in ThriveStack citedby.
ThriveStack citedby groups your tracked prompts by topic, shows visibility per topic and engine, and suggests new topics and prompts to fill the gaps in your map.