Track any prompt you want in the Answer Engine Monitor
Until now the monitor watched a fixed set of category prompts. You can now write your own — up to 50 per location — and the agent runs them nightly across ChatGPT, Perplexity, Gemini and Google AI Overviews, storing every answer verbatim so you can see exactly which sentence cited a competitor instead of you.
Tracked prompts
12 of 50 usedbest coffee shop near Ballard for remote workChatGPT#2
where to get oat milk latte in SeattlePerplexity#1
coffee shops open before 6am SeattleGemini#4
quiet cafe with wifi FremontAI Overviews#6
dog friendly coffee patio SeattleChatGPT#3
specialty espresso roasters near mePerplexity#8
Share of answer
30 days
+14 pts vs. June
2,140 answers stored
New
- Custom prompts — write up to 50 per location; the agent runs them nightly and keeps the raw answer text.
- Answer diffing — see exactly what changed in an engine's answer between two runs.
- Competitor watchlist — pin up to 5 rivals and get share-of-answer for each in the same view.
Improved
- Monitor runs finish ~4× faster — a 200-location account now completes in under 9 minutes.
- Citation links resolve to the exact source page instead of the domain root.
- CSV export includes the full answer body, not just position.
Fixed
- Prompts containing an apostrophe were silently dropped from nightly runs.
- Gemini results occasionally recorded under the wrong location in multi-location accounts.
- Share-of-answer chart rendered 0% during the first 24 hours after adding a prompt.
- The date picker allowed a range ending before it started.
- Long prompts overflowed the row on narrow viewports.
- Email digest linked to the previous week's report.
- Duplicate prompts could be added if pasted with trailing whitespace.
- Sorting by position put unranked prompts first instead of last.
8 fixes