Enhanced · live dataAI Visibility Agent

AI Answer Gap

Find the prompts where AI should cite your brand but cites a competitor instead, ranked by AI search demand so you fix the biggest gaps first.

Copy run promptView on GitHubInstall

Agent-native

Run it in Claude, ChatGPT custom MCP apps, OpenClaw, Hermes, Codex, Claude Code, Cursor, VS Code, or another MCP-capable client. No dedicated GUI flow and no separate LLM API key.

Backed by live public data

Every step is grounded in live public-data records UnifAPI returns, so the output cites what is actually ranking, posting, or being said — not a generic best-practice list.

Composable & open source

Skills cross-reference each other and live in a public, MIT-style repo. Read the full SKILL.md on GitHub, fork it, or run it as-is inside your agent.

Run prompt

Paste this into Codex or Claude Code

The prompt is intentionally editable. Replace the handles, market, budget, and campaign goal, then let the agent call UnifAPI MCP when it needs live public data.

Find the AI-answer prompts we should own but don't for our category, name who owns each answer instead, and rank them by AI search volume into a content backlist with quick wins flagged.
How the skill works

The full skill, rendered from its SKILL.md

Turn observed AI-answer gaps into a prioritized research and content backlog. Use an existing ai-visibility-audit or collect a dated panel first. Read available product context so proposed topics match what the product can credibly help with.

Use UnifAPI for live evidence

Connect with the unifapi skill to discover current schemas, prices and authentication before collecting data.

Workflow

  1. Select relevant prompts. Group them by use case and search intent. /geo/keywords/search-volume supplies optional demand estimates; an absent estimate is unknown, not evidence that a useful niche prompt has no value.
  2. Check current answers. Use /geo/answers for ChatGPT/Gemini, /geo/serp for Google AI Mode, and /seo/serp with include_ai_overview for Google Search AI Overviews. Keep engines, markets and natural/forced-search settings separate. Preserve exact answers and cited source URLs.
  3. Describe the observed gap. Distinguish no answer, unsuccessful collection, brand absent from this answer, name-only mention, and cited brand. A sample is not proof of universal absence. Inspect references; top-level is_target and unused search results do not by themselves establish a citation.
  4. Research sources. Open the actual cited pages. /geo/mentions/top-pages, /top-domains, and /search add separate indexed-corpus context. They do not supply the winning URL for a particular live prompt. /cross-aggregated-metrics gives group counts, not ownership of your prompt-panel gaps.
  5. Propose a response. Identify the missing information, evidence or third-party representation. Choose an existing-page update, a new useful page, or an external research lead. Each proposed change needs a concrete source and a plausible benefit to readers. Name-only mentions and existing organic rankings are clues, not proof that formatting is the cause.
  6. Prioritize. Consider product relevance, evidence of demand, gap quality, available expertise, and effort. Show the inputs and uncertainty. Optional scoring is a planning aid, not a predicted citation lift.
  7. Validate later. Use the same panel and multiple dated runs after changes. Compare paired successful samples, show completion/answer rates, and avoid attributing every change to the edit.

Output

For each priority, give the exact prompt, engine/configuration, observed brand status, cited competitors or sources, suggested page/action, evidence, effort estimate and a validation plan. Preserve request ids and billed credits. An API failure should appear as missing evidence and never as a content gap.

See the audit methodology for coverage/share definitions, citation counting and evidence boundaries. Use llm-mention-tracking for budgeted snapshots and changes over time. This skill researches and drafts; it does not publish content or send outreach without authorization.

Source: ai-answer-gap/SKILL.md on GitHub — open a PR there to improve it.

Public-data tools

The live APIs this skill calls

Every operation the skill names is one of these UnifAPI platforms — still visible and callable for product code, debugging, and custom agent flows.

  • Prioritized gap backlist: prompt, AI search volume, the owner, and the cited source
  • Per gap: the winning angle to beat and a quick/net-new tag
  • Fast wins where the brand ranks organically but isn't cited
  • Every gap cited to the AI answer and dated

GEO

AI answer visibility, LLM mentions, AI search volume

SEO

Keyword volume, SERP checks, backlinks, domain rank

Related skills

More skills in the AI Visibility Agent

Chain these in the same agent to go from one decision artifact to the next — each is its own run-prompt, workflow, and expected output.

AI visibility audit

The GEO equivalent of an SEO audit — check whether you're cited in AI answers and who owns the answer instead.

Open AI visibility audit

AI visibility fixes

Prioritize Structure, Authority, and Presence fixes for prompts where AI answers do not cite you.

Open AI visibility fixes

LLM mention tracking

Track how often you're mentioned across ChatGPT and AI search over a fixed prompt set, and how share of voice moves vs competitors.

Open LLM mention tracking
See every skill in the AI Visibility Agent