SkillsAEO Keyword Research and Prompt Simulation: Agent Skill
AEO Keyword Research and Prompt Simulation: Agent Skill
Turn a brand's real search demand into the prompts to track in AI answers: keyword research, a prompt set and a first share-of-voice baseline, in one workbook.
Get full access to this skill and the tools to run it
For agents
Your AI gets this skill, your clients' context and the research tools, in Claude or any MCP client.
Connector URL: https://mcp.unilytics.ai/mcp
What you get
A checked workbook: the questions to track in AI answers and the keywords behind them, each tied to a page, with likely questions written where real search data runs out (the prompt simulation). Then a first measurement of how often AI answers name the brand, and its share of voice.
Deliverable: Workbook
When to use it
Keyword research for a client, or building the set of prompts to track in AI answers.
The steps
The skill takes your agent through these steps, in order. Each step's full instructions come with the skill when you run it through Unilytics.
- Scope the product or service (before any keyword)
- Category-anchored keyword research
- The prompt universe
- Baseline and re-score
The full skill also covers: File map · Before you start · The anchoring principle · Never silently degrade a blocked data path · Deliverable · Definition of done · Guardrails · Rules for this skill.
Tools it needs, with or without Unilytics
Straight from the skill: the data each step needs and the tools that can supply it. Unilytics is one option, never a requirement.
| Capability | Ways to get it |
|---|---|
| Keyword volume, CPC, competition, 12-month trend | Keywords Everywhere, Semrush, Ahrefs or DataForSEO; or the Unilytics MCP keyword_research |
| Keyword difficulty, keyword gap, bulk overview, phrase-match expansion, rankings by URL | Semrush or Ahrefs (UI export or API) |
| Live SERP composition, People Also Ask, AI Overview presence, did-you-mean | DataForSEO SERP API or the live Google results page; or the Unilytics MCP serp_results and aeo_serp_urls |
| Competitor page meta, titles, H1/H2, FAQ blocks | Firecrawl or a browser; or the Unilytics MCP content_scrape, batch_scrape, js_scrape |
| Backlink anchor text | Ahrefs, Semrush or Moz; or the Unilytics MCP backlink_analysis |
| Category trend direction | Google Trends; or the Unilytics MCP google_trends |
| Competitor discovery and overlap | Semrush or Ahrefs competitor reports |
| Search Console queries and pages | Your own Search Console connector or export, or its bulk export to BigQuery (which data/longtail_split_bigquery.sql reads) |
| Engine grounding queries | Bing Webmaster Tools AI performance report (export) |
| Observed question data | AlsoAsked, Answer Socrates, AnswerThePublic, Google People Also Ask and related searches |
| Paid copy of competitors | Google Ads Transparency Center (browser; APIs that return only creative ids carry no copy) |
| Analyst and review categories | Gartner Peer Insights, G2, Capterra (browser) |
| AI citations and assistant searches for the cross-check | Google AI Mode, Perplexity and ChatGPT, with web search on |
| Prompt drafting, filtering, volume estimate, clustering | The method above, by hand; optionally the Unilytics MCP aeo_generate_prompts, aeo_filter_prompts, aeo_prompt_volume, aeo_cluster_prompts, aeo_cluster_by_urls, aeo_audit_prompts |
| Brand visibility baseline across engines | A prompt-tracking tool or the engines directly; or the Unilytics MCP brand_visibility |
| Live URL status and site page list | A crawler such as Screaming Frog; or the Unilytics MCP site_map, aeo_site_pages, site_crawl |
| Web research (category, launch posts, docs indexes) | A search engine; or the Unilytics MCP web_research_tool |
| Prior research and project settings | The client's research files; or the Unilytics MCP kb_search and get_project_config |
| Build the workbook | Any spreadsheet tool; in Python, openpyxl. The gate scripts in data/ need only the standard library |
Without the Unilytics MCP
Every step runs without Unilytics. Without it, what you lose is one connection for several of the capabilities above: each one then comes from the other ways listed for it (a tool or account you already have, a vendor API, or a manual export). The Unilytics MCP is one optional way to get several of these tools in a single connection: https://unilytics.ai
New accounts start with 50 credits. Connecting, and reading your skills, projects and client context, costs nothing; research tools and some writes use credits.