Guides

MCP recipes for GEO teams

Questions and workflows GEO teams run through OClever every week. Copy the request into your assistant and adjust names, dates and engines.

Monday visibility pulse

For: Marketing leads · Needs: any paid plan

Ask your assistant
Using OClever, compare our AI visibility for the last 7 days with the 7 days before. Show visibility score, share of voice, average position and citation share, the two engines and two topics that moved most, and any competitor that gained more than 3 points. Keep it to a short table and three bullets.

Tools used: get_visibility_overview, get_engine_report, get_topic_report, get_competitor_comparison

Or run the weekly_visibility_pulse prompt. Paste the result into your team channel.

Why did a competitor jump?

For: GEO and SEO teams · Needs: any paid plan

Ask your assistant
Acme's share of voice went up this month. Which prompts and engines drove it, which sources cited Acme in those answers, and did any of those sources stop citing us?

Tools used: get_metric_report, get_prompt_results, get_citation_sources, get_cited_pages

Ask for get_metric_report grouped by prompt for the competitor, then drill into the answers.

Citation gap list for outreach

For: PR and content teams · Needs: any paid plan

Ask your assistant
List the pages cited for our competitors but never for us in the last 3 months, grouped by domain type (reviews, listicles, forums, news). For each domain, give the number of answers it appeared in and the competitors it cited.

Tools used: get_cited_pages, get_citation_sources

Uses gap_only. The citation_gap_plan prompt adds a suggested action per source.

Prompts where we never appear

For: Content strategists · Needs: any paid plan

Ask your assistant
Which active prompts have we not been mentioned in for 4 weeks? Who wins each one, and what do the winning answers have in common?

Tools used: get_prompt_results, get_prompt_answers, get_answer

Set diagnosis to absent. Ask the assistant to quote the winning answers' sources.

Turn a gap into a content brief

For: Content strategists · Needs: any paid plan

Ask your assistant
For the prompt “best payroll software for startups”, read the latest answers from ChatGPT and Perplexity, list the claims and sources they rely on, and write a brief for a page that could earn a citation: target question, headings, proof points and sources to reference.

Tools used: list_prompts, get_prompt_answers, get_answer

Read-only. To have OClever write the draft, follow with the next recipe.

Draft the article in Content Studio

For: Content teams · Needs: Growth and above, spend scope

Ask your assistant
Draft an article for that prompt with two images. Show me the coin cost first.

Tools used: draft_article, get_article_draft

Costs 40 coins plus images. The draft stays unpublished in Content Studio for review.

Negative claims to fix

For: Brand and product marketing · Needs: any paid plan

Ask your assistant
What are the most common negative or inaccurate things AI answers say about us this quarter, which engines say them, and which sources do those answers cite?

Tools used: get_sentiment_report, get_prompt_answers, get_citation_sources

Pair each claim with the page on your site that should correct it.

Prompts for a product launch

For: Product marketing · Needs: Growth and above, write scope

Ask your assistant
We are launching an expense-management feature. Draft 15 prompts buyers would ask an AI assistant across awareness, consideration and decision, put them in a new topic called Expenses, and show me the preview before saving.

Tools used: list_topics, list_prompts, create_prompts

Free, counts against your prompt quota. The preview flags duplicates of existing prompts.

Check a fix landed

For: SEO teams · Needs: Growth and above, spend scope

Ask your assistant
We updated our pricing page yesterday. Re-run our five pricing prompts on ChatGPT and Perplexity now and tell me whether we are mentioned and cited.

Tools used: list_prompts, run_collection_now, get_prompt_answers

5 prompts × 2 engines = 10 coins. Single runs are noisy; confirm the trend over the following weeks.

Test a question live

For: Anyone · Needs: Scale and above, spend scope

Ask your assistant
Ask ChatGPT, Perplexity and Gemini right now: “Which AEO platforms are best for agencies managing many clients?” Compare which brands each one names and which sources they cite.

Tools used: run_prompt_lab, get_prompt_lab_run

10 coins per engine, so 30 here. Nothing is added to tracking.

Triage recommendations

For: Team leads · Needs: Growth and above, write scope

Ask your assistant
List open recommendations with high impact and low effort. Approve the top three and dismiss anything about press releases with the note “not a channel for us”.

Tools used: get_recommendations, update_recommendation

Free. Each update previews first.

Agency roll-up across clients

For: Agencies · Needs: any paid plan

Ask your assistant
For each workspace I can see, give visibility score and share of voice for the last 4 weeks with the change, and flag any client that dropped by more than 5 points.

Tools used: list_workspaces, get_visibility_overview

Connect with all client workspaces selected. Each client login only ever sees its own workspace.

Are AI crawlers reading the new pages?

For: Technical SEO · Needs: any paid plan

Ask your assistant
Which AI crawlers visited our site in the last 7 days, which pages did they visit most, and did GPTBot or PerplexityBot fetch anything under /guides/?

Tools used: get_crawler_activity

Needs server or CDN logs connected in OClever.

Agent readiness fix list

For: Web teams · Needs: Scale and above

Ask your assistant
Summarise our latest agent readiness result and turn the failed tasks into a fix list ordered by impact, with the page and change for each.

Tools used: get_agent_readiness

Or run the agent_readiness_fix_plan prompt.

Last updated 5 October 2026. Questions or something unclear? Email [email protected].