ThriveStack citedby Prompt Research turns your brand kit, topics, personas, competitors and markets into a tagged prompt panel of real buyer questions. You edit the panel, then pick what to track.
Step through how ThriveStack citedby researches real buyer questions from brand signals, tags every prompt, and tracks AI citations.
A prompt panel is a fixed set of real buyer questions you run through ChatGPT, Perplexity, Gemini and AI Overviews on a schedule. It measures how often AI names your brand across answer engine optimization (AEO) and GEO.
AI engines publish no query data, so the questions you put in your panel decide what your visibility score means. ThriveStack citedby Prompt Research builds that panel from your brand signals. You edit it, then choose what to track.
You give it a starting point. It returns a tagged panel. You decide what runs. Most teams get a first panel live in an afternoon.
Enter your domain, a few keywords or a list of topics. Add your personas, competitors, markets and languages.
Each prompt is a long, real question tagged by topic, persona, intent, market and branded or not. Edit any of them.
Recommended prompts are flagged. Pick what to track. Each one then runs about seven times a day on every engine.
This is the prompt panel inside ThriveStack citedby. Prompts sit under topics. Every row shows visibility, position and the brands AI named. You track the ones that matter and archive the rest.
Topics keep the panel tidy. Four topics hold the 18 prompts. Click a topic to see just its prompts, or add a new one.
The tracker counts what you run. 18 / 25 means 18 prompts are tracked out of 25 slots. Suggested prompts wait in their own tab until you add them.
Twelve signals shape a good panel. Each one adds a tag to every prompt, and the tags let you slice results later.
Skip a signal and you get a blind spot. Leave out markets and you never learn you are invisible in Germany.
One blended score hides the truth. Tags turn it into answers you can act on, such as "we win CFO prompts in the US but vanish in Germany."
Report by engine and by market, never as one number. Cross-engine gaps are normal. ThriveStack research found only 11% source overlap between ChatGPT and Perplexity.
| # | Metric | Cut it by |
|---|---|---|
| 01 | Mention rate | Engine, intent, persona |
| 02 | Share of voice (Share of Model) | Topic, market, competitor |
| 03 | Citation rate | Engine, topic, source domain |
| 04 | Sentiment and accuracy | Branded prompts only |
| 05 | Top cited sources | Topic, engine |
Multiply topics by 5 to 10 prompts each, then by personas, markets and languages, and buyer intents. Then multiply by the AI engines you track.
A startup lands near 45 prompts, a growing brand near 360, and an enterprise can pass 10,000.
3 topics x 5 prompts x 1 persona x 1 market x 3 intents. About 135 prompt and engine pairs on 3 engines.
5 topics x 6 prompts x 2 personas x 2 markets x 3 intents. About 1,440 pairs on 4 engines.
More topics, more variants, more markets and languages, on more engines.
Prompt Research is the first step. The panel you build feeds the rest of ThriveStack citedby.
Each prompt runs about seven times a day on every engine, so the numbers carry real signal.
Mention rate, share of voice, citation rate and sentiment, each cut by the tags you set.
When rivals get quoted and you do not, ThriveStack citedby points to the owned, earned and technical fixes to make.
Prompt Research finds the questions your buyers ask AI engines and turns them into a tagged prompt panel you can track. You start from your domain, keywords or topics. It returns prompts tagged by topic, persona, intent, market and branded or not. You edit the panel, then pick the prompts to run.
It reads your brand kit, topics, personas, competitors, markets, languages and intent mix in one pass. Each prompt is a long, real buyer question. Every prompt carries tags, so you can slice results later. You can edit any prompt before it runs.
No. OpenAI, Google and Perplexity do not publish what people type, so no tool can read private chats. Any prompt volume you see anywhere is an estimate. Prompt Research starts from your own brand signals and the demand data you already hold, such as Search Console queries and sales call notes.
Keyword research gives you short phrases and a rank. AI search works another way. A buyer types a long question and the engine writes a fresh answer each time. So you track a sample of real questions and report how often AI names you. Rank order almost never repeats.
No. Prompt Research flags the prompts it recommends, and you choose which ones to track. Start wide, then trim. Let new prompts run for at least two weeks before you cut any. Keep the core panel stable for a quarter so trends stay comparable.
Multiply topics by 5 to 10 prompts each, then by personas, markets and languages, and buyer intents. A startup lands near 45 prompts. A growing brand lands near 360. An enterprise can pass 10,000. Each prompt then runs about seven times a day on every engine.
Sign up for ThriveStack citedby and start from your domain, keywords or topics. You can also book free onboarding and we set up your first panel with you.
Start from your domain. Edit every prompt before it runs. Pick the ones to track and see where AI names you.