MCP server
Example prompts
Copy-paste questions that get great answers from the Analyze AI MCP server — from one-line status checks to full competitive playbooks.
Built-in playbooks
Seven workflows ship with the server. Ask for them by name or by intent — your assistant runs the whole recipe and returns a finished deliverable with evidence:
| Playbook | What you get |
|---|---|
| Weekly visibility brief | Where you stand per engine, what changed and why, sentiment watch, and 2–3 next moves — under 400 words |
| Why did visibility change? | Root-cause analysis: which prompts, engines, and competitors drove a move |
| Competitor radar | Who's gaining, where they outrank you, and which sources power them |
| Sentiment deep dive | What AI models literally say about your brand — quoted spans, drivers, and fixes |
| Citation gap plan | Ranked outreach targets: the pages and domains to earn mentions on, with rationale |
| Campaign before / after | Did a launch move AI visibility? Windows compared either side of your date |
| Turn gap into brief | Your biggest visibility gap converted into a content brief grounded in your proof points |
"Run the weekly visibility brief."
"Did our July 2 launch move AI visibility? Compare 30 days each side."
Status & trends
"I want visibility over the past 14 days, and sentiments, and a perception map."
"How did our visibility move this week compared to the month? Chart it."
"Which prompts had the biggest visibility swings in the last 30 days?"
"Which AI engines are weakest for us right now?"
Competitive intelligence
"Who mentions [competitor] in AI answers without mentioning us?"
"Which competitor is gaining visibility the fastest, and on which prompts?"
"Compare us against [competitor] on visibility, sentiment, and citations."
"What part of the narrative do we own, and which does [competitor] own?"
Sources & citations
"Show me citation types over the last 30 days and the highest-cited URLs."
"Which high-authority pages cite competitors but not us? Rank by authority."
"Is [url] worth chasing a mention on?"
"Are any risky sources citing us?"
Perception & sentiment
"What does AI actually say about our brand? Quote the exact lines."
"Where is our sentiment lowest, and what's driving it?"
"Which of our brand themes does AI under-represent or frame negatively?"
Prompt management (write access)
"Add these five prompts, set them to the US, and run them now."
"Which tracked prompts have had zero mentions in 90 days? Retire them."
"Track the suggested prompts about [topic] and tell me when the first results land."
Getting better answers
Ask for receipts. Add "quote the evidence" to any question and the assistant will back every claim with the exact span from a real AI answer — and tell you when a number has no evidence behind it.
- Name a window. The strategy analyses are pre-computed for 7, 14, 30, and 90 days — "over the last 30 days" gets richer answers than "recently".
- Ask for the chart. "Chart it as a line/bar/pie" works in chart-capable assistants; the data already knows its preferred shape.
- Chain questions. The best sessions drill down: brief → "why did engine X drop?" → "who took the position?" → "turn that gap into a content brief."