These work in any connected client. Name the site by domain when you track more than one.
Understand where you stand
- Which of my tracked prompts lost citations this month, and which page should I fix first?
- What is my citation rate on Perplexity versus ChatGPT for example.com over the last 30 days?
- Which competitors show up most often in the answers where we do not?
- Show me the answers where we are mentioned in the text but not cited in the sources.
Find the work
- List my high-priority missions for example.com and summarise the three that share a root cause.
- Which domains does ChatGPT cite in my category, and do any of them mention us?
- Suggest ten prompts I am not tracking yet, weighted toward comparison questions.
- Are there factual errors in what AI answers say about us right now?
Do the work
- Add five prompts about serverless database pricing and check them now.
- I published the new comparison page today. Log it against the prompts it targets.
- Mark the mission for the pricing-page prompt as in progress.
- Turn off the four prompts that have never returned a citation in six months.
Close the loop
- Did the page I shipped three weeks ago move anything yet?
- Read me the verdict for the last content change, including the evidence grade.
- How much of my monthly check quota is left, and what would ten more prompts cost me?
Ask the agent to check usage before a batch of work. It can plan around the remaining quota instead of hitting the limit halfway through.
Chaining
The tools are designed to be combined. A useful pattern in a coding agent: read the missions for a site, open the page the mission points at, edit it, then log the content change in the same session so the before-and-after window opens the moment you deploy.