Not all prompts perform equally. Prompt Performance breaks down your mention rate, trend, and best/worst engine for every prompt in your library — so you fix the right things first.
Start analyzing prompts →Mention rate
What percentage of scans for this prompt resulted in a brand mention. The primary performance indicator.
Average score
Average MentionShare score across all scans that used this prompt.
Best engine
Which AI engine mentions your brand most reliably for this specific prompt.
Worst engine
Which engine never or rarely mentions you. The highest-value optimization target.
30-day trend
Whether mention rate for this prompt is trending up, flat, or down over the last 30 days.
Last scanned
When this prompt was last included in a scan. Helps identify stale prompts that need re-testing.
TrueCite automatically classifies every prompt by intent using Claude Haiku. Filter by intent to focus your optimization effort on the highest-value prompts.
Example prompt
“"What is the best HR onboarding software for mid-market?"”
High-value. Buyers are comparing tools. Being cited here drives pipeline.
Example prompt
“"How does AI-powered HR onboarding work?"”
Top of funnel. Cited here builds brand authority before the purchase decision.
Example prompt
“"[Your Brand] HR onboarding pricing"”
Bottom of funnel. If AI doesn't cite you here, you may be losing late-stage buyers.
Example prompt
“"Sign up for [Your Brand] HR software"”
High-intent. Being cited here directly converts to trials and demos.
TrueCite classifies every prompt by buyer intent. Here is what each type means and why it matters.
“"Best [tool] for [use case]?"”
Buyer is evaluating. Being cited here drives shortlisting and trials.
Highest priority
“"What is [concept]?"”
Buyer is researching. Cited here builds authority before purchase decision.
Awareness stage
“"[Brand] pricing"”
Buyer is looking for you specifically. Must appear here at any cost.
Bottom funnel
“"Sign up for [tool]"”
Buyer is ready to act. Highest conversion intent of all prompt types.
Direct conversion
Sort by lowest mention rate
Find your most broken prompts — the ones where AI engines never recommend you despite the prompt being core to your category.
Filter by commercial intent
Focus optimization effort on prompts that directly map to purchase decisions.
Compare best vs worst engine per prompt
If ChatGPT cites you but Perplexity never does, run a targeted gap analysis against Perplexity's preferred sources.
Check the 30-day trend after publishing fixes
Validate whether your FAQ blocks and schema changes are moving the needle on specific prompts.
Prompt Performance is available on Starter and Pro plans. Start your 7-day free trial to explore your MentionShare score first, then upgrade to unlock Prompt Performance.
Start free trial →It shows your mention rate for each individual prompt rather than one blended score. That distinction matters because performance varies enormously between questions — a brand can be named in nearly every answer to one question and absent from a closely related one. Seeing it per prompt tells you which questions to write content for.
The Performance tab requires Starter or above. Trial and Lite accounts can still create and scan prompts and see results per scan; the per-prompt performance history, engine breakdown and trend views are the part gated to Starter and up.
Trial runs 1 prompt per scan, Lite 5, Starter 20, and Pro and Enterprise 30. Separately, each plan has a monthly prompt budget that every scan draws from — 150 a month on Lite, 500 on Starter, 2,000 on Pro and 10,000 on Enterprise at standard regional pricing. One prompt covers every engine and country you have selected, so adding engines does not increase what a scan costs you.
Yes, on Starter and above. Connect Google Search Console and TrueCite suggests your existing search queries as prompts, showing clicks, impressions, CTR and position for each so you can pick the ones worth tracking. Imported queries are classified by intent automatically, and queries already in your library are filtered out.
It labels each prompt as commercial, informational, navigational or transactional so you can separate buying questions from research questions. A brand often performs well on informational prompts and poorly on commercial ones, which is exactly the gap worth fixing first — filtering performance by intent makes that visible instead of averaging it away.