AEO & GEO skill 07 of 13

Perplexity, Copilot, Claude and Gemini Visibility SKILL.md

The Perplexity, Copilot, Claude and Gemini Visibility skill makes Claude, Codex or another AI agent improve visibility engine by engine. It maps each engine’s retrieval source and documented crawlers, uses Bing Webmaster Tools AI Performance, the only first-party citation data available, checks IndexNow, Google-Extended and Claude crawler decisions, and diagnoses why one engine mentions you while another doesn’t.

answer-engine-platforms/SKILL.md, version 1.0.0, 1,156 words.

Each answer engine retrieves from a different place, credits sources differently and gives site owners different data. Perplexity cites heavily and visibly. Microsoft Copilot is built on Bing and is the only engine that gives site owners citation counts. Claude searches the web when a question needs it. The Gemini app grounds answers with Google. A plan that works for one may miss another.

This skill audits and improves visibility engine by engine, using each platform's documented crawlers and reports.

Engine reference

EngineMain retrieval sourceCrawlersSite owner dataNotes
PerplexityIts own index plus other sourcesPerplexityBot (index, respects robots.txt); Perplexity-User (user-triggered, generally ignores robots.txt per Perplexity's documentation) [Confirmed]None beyond referralsShows many citations per answer; strong on fresh and niche topics [Consensus]
Microsoft Copilot and Bing AI summariesBing indexBingbotBing Webmaster Tools AI Performance: citations, cited pages, grounding queries, citation share [Confirmed]Whoever is strong in Bing is well placed for Copilot [Consensus]
Claude (web search on)Web search providers plus fetchingClaude-SearchBot (search), Claude-User (user-triggered), ClaudeBot (training) [Confirmed by Anthropic]None beyond referralsProvider details aren't fully public [Speculation on specifics]
Gemini appGoogle Search groundingGooglebot; Google-Extended token controls training use and grounding in Gemini Apps and Vertex AI [Confirmed, Google crawler docs]None specificBlocking Google-Extended can limit Gemini app use of your content, not Google Search

Operating principles

  1. One access policy per engine, written down. Search and user-triggered fetchers allowed; training crawlers a business decision.
  2. Bing is a multiplier. Copilot and several partners rely on it; Bing Webmaster Tools and IndexNow are low effort.
  3. Use the only citation data that exists. Bing AI Performance shows which pages and grounding queries earn citations; learn from them.
  4. Freshness matters more on some engines. Perplexity and Copilot often favour recently updated sources for time-sensitive questions. [Consensus]
  5. Test each engine separately. Don't assume a mention in one means a mention in all.

When to use, and when not to

Use it when one engine ignores you while others don't, when B2B buyers use Copilot inside Microsoft 365, when Perplexity referrals matter, or when deciding on Google-Extended and Claude crawler policies.

Don't use it for ChatGPT (chatgpt-search-visibility) or Google AI Overviews and AI Mode (ai-overviews-ai-mode).

Inputs to ask for, in one message

  1. robots.txt, CDN bot rules and 30 days of logs.
  2. Bing Webmaster Tools access (and whether IndexNow is on).
  3. The prompt set and markets.
  4. GA4 referrals from perplexity.ai, copilot.microsoft.com, bing.com, claude.ai, gemini.google.com.
  5. The business's stance on AI training crawlers.

Raw SKILL.md file

.md

This is the exact file agents read, including the credit header and licence notice. Save it as answer-engine-platforms/SKILL.md.

---
name: answer-engine-platforms
description: "Improve visibility engine by engine in Perplexity, Microsoft Copilot and Bing AI summaries, Claude and the Gemini app: each engine's retrieval source and documented crawlers, Bing Webmaster Tools AI Performance citation data, IndexNow, Google-Extended and Claude crawler decisions, per-engine prompt testing and diagnosis. Use whenever someone asks how to appear in Perplexity, Copilot, Bing AI, Claude or Gemini answers, about PerplexityBot, ClaudeBot or Google-Extended, Bing AI Performance citations, or why one AI engine mentions them and another doesn't."
license: CC-BY-4.0
metadata:
  author: Afeez Adebayo
  author_url: https://afeezadebayo.com
  author_linkedin: https://www.linkedin.com/in/afeez-adebayo/
  author_x: https://x.com/za_icon
  source: https://afeezadebayo.com/skills-md/marketing/aeo-geo/answer-engine-platforms/
  category: marketing
  subcategory: aeo-geo
  version: "1.0.0"
  created: "2026-10-10"
  updated: "2026-10-10"
---

# Perplexity, Copilot, Claude and Gemini Visibility

Each answer engine retrieves from a different place, credits sources differently and gives site owners different data. Perplexity cites heavily and visibly. Microsoft Copilot is built on Bing and is the only engine that gives site owners citation counts. Claude searches the web when a question needs it. The Gemini app grounds answers with Google. A plan that works for one may miss another.

This skill audits and improves visibility engine by engine, using each platform's documented crawlers and reports.

## Engine reference
| Engine | Main retrieval source | Crawlers | Site owner data | Notes |
|---|---|---|---|---|
| Perplexity | Its own index plus other sources | PerplexityBot (index, respects robots.txt); Perplexity-User (user-triggered, generally ignores robots.txt per Perplexity's documentation) [Confirmed] | None beyond referrals | Shows many citations per answer; strong on fresh and niche topics [Consensus] |
| Microsoft Copilot and Bing AI summaries | Bing index | Bingbot | Bing Webmaster Tools AI Performance: citations, cited pages, grounding queries, citation share [Confirmed] | Whoever is strong in Bing is well placed for Copilot [Consensus] |
| Claude (web search on) | Web search providers plus fetching | Claude-SearchBot (search), Claude-User (user-triggered), ClaudeBot (training) [Confirmed by Anthropic] | None beyond referrals | Provider details aren't fully public [Speculation on specifics] |
| Gemini app | Google Search grounding | Googlebot; Google-Extended token controls training use and grounding in Gemini Apps and Vertex AI [Confirmed, Google crawler docs] | None specific | Blocking Google-Extended can limit Gemini app use of your content, not Google Search |

## Operating principles
1. **One access policy per engine, written down.** Search and user-triggered fetchers allowed; training crawlers a business decision.
2. **Bing is a multiplier.** Copilot and several partners rely on it; Bing Webmaster Tools and IndexNow are low effort.
3. **Use the only citation data that exists.** Bing AI Performance shows which pages and grounding queries earn citations; learn from them.
4. **Freshness matters more on some engines.** Perplexity and Copilot often favour recently updated sources for time-sensitive questions. [Consensus]
5. **Test each engine separately.** Don't assume a mention in one means a mention in all.

## When to use, and when not to
Use it when one engine ignores you while others don't, when B2B buyers use Copilot inside Microsoft 365, when Perplexity referrals matter, or when deciding on Google-Extended and Claude crawler policies.

Don't use it for ChatGPT (`chatgpt-search-visibility`) or Google AI Overviews and AI Mode (`ai-overviews-ai-mode`).

## Inputs to ask for, in one message
1. robots.txt, CDN bot rules and 30 days of logs.
2. Bing Webmaster Tools access (and whether IndexNow is on).
3. The prompt set and markets.
4. GA4 referrals from perplexity.ai, copilot.microsoft.com, bing.com, claude.ai, gemini.google.com.
5. The business's stance on AI training crawlers.

## Workflow

### Step 1. Access by engine
Check each crawler in robots.txt and logs; verify by IP where providers publish ranges. Remove CDN challenges for search and user-triggered fetchers.

### Step 2. Bing foundations
Verify the site in Bing Webmaster Tools (it can import from Search Console), submit sitemaps, turn on IndexNow, fix crawl errors. Check that key pages are indexed in Bing.

### Step 3. Read Bing AI Performance
List pages with the most citations and their grounding queries. Note citation share for important queries. Study what cited pages have in common: structure, freshness, data, comparisons.

### Step 4. Run the prompts per engine
Record mentions, citations, accuracy and competitors for Perplexity, Copilot, Claude (search on) and Gemini.

### Step 5. Diagnose per engine
| Pattern | Likely cause | Action |
|---|---|---|
| Missing only in Copilot | Weak Bing indexing or ranking | Bing foundations, IndexNow, Bing-specific checks |
| Missing only in Perplexity | PerplexityBot blocked, or no fresh, specific pages | Access, freshness, specific data |
| Missing in Claude | Claude-SearchBot or Claude-User blocked, or weak results for the searches run | Access, answer-ready pages |
| Missing in Gemini app but present in Google | Google-Extended blocked | Review the token decision |

### Step 6. Improve and re-test
Apply fixes, then re-run the engine's prompts after two to four weeks.

## Copilot inside work tools
Many business buyers meet Copilot in Microsoft 365 and Edge, not on bing.com. For B2B brands, Copilot visibility can matter more than consumer traffic data suggests.

## Worked example
A B2B HR platform was cited in ChatGPT and Perplexity but never in Copilot. Bing Webmaster Tools showed 40% of key pages not indexed after a migration, and IndexNow wasn't set up. After resubmitting sitemaps, fixing redirect chains Bing still crawled and enabling IndexNow, AI Performance began showing citations for comparison and pricing pages within five weeks, and Copilot named the brand for 7 of 15 tracked prompts.

## Output format
```
# Answer engine review: <brand>, <date>
## Access matrix (engine x crawler: allowed, verified, issues)
## Bing foundations and AI Performance summary
## Prompt results by engine
## Diagnosis and actions per engine, with re-test dates
```

## Testing each engine
| Engine | Test setting | What to record |
|---|---|---|
| Perplexity | Default search, fresh session, target country | Cited sources and their order |
| Copilot | Copilot on web and, for B2B, inside Microsoft 365 or Edge | Mentions and linked sources |
| Claude | Web search enabled | Whether it searches, sources fetched, description accuracy |
| Gemini app | Default settings | Mentions, links and whether Google sources appear |

## Engine checklist
| Engine | Must-have checks |
|---|---|
| Copilot | Bing indexing, sitemaps, IndexNow, Bing AI Performance citations |
| Perplexity | PerplexityBot allowed at robots.txt and edge; fresh, specific pages |
| Claude | Claude-SearchBot and Claude-User allowed; facts in server HTML |
| Gemini app | Google-Extended decision documented; strong Google visibility |
Run the checklist quarterly and after any CDN, robots.txt or hosting change, because these engines can lose access silently.

Log every engine change you make, with the date, so later shifts in citations can be explained.

## Edge cases and judgement calls
| Situation | What to do |
|---|---|
| Bing indexes far fewer pages than Google | Fix redirects, sitemaps and quality signals Bing reports; use IndexNow |
| Perplexity cites forums over your guides | Add first-hand detail and specifics that forums offer; join the conversation with disclosure |
| Claude never fetches the site | Check Claude-SearchBot and Claude-User access at the edge |
| Gemini app ignores you but Google ranks you | Check whether Google-Extended is blocked and whether that was intended |
| A partner integration cites you heavily in Bing AI Performance | Strengthen the cited pages and keep them current |

## Handoffs to related skills
- Crawler policy: `ai-crawler-access`
- Page improvements: `answer-ready-content`
- Reporting: `ai-visibility-tracking`
- Missing recommendations: `ai-citation-earning`
- Bing crawl issues: the Technical SEO `redirects-status-codes` and `xml-sitemaps` skills

## Quality bar: what an excellent review includes
- An access matrix per engine and crawler.
- Bing foundations checked and AI Performance read.
- Diagnoses specific to each engine.
- Re-test dates per engine.

## Common mistakes
- Treating all AI engines as one.
- Ignoring Bing because its search share looks small.
- Blocking Google-Extended without knowing it affects the Gemini app.
- Blocking user-triggered fetchers that represent real people.

## Questions to ask before you recommend changes
- Which engines do your buyers use at work and at home?
- Is AI training access acceptable to the business?
- Is Bing Webmaster Tools set up and monitored?

## Self-check before you answer
- Did I check access per crawler and verify by IP where possible?
- Did I use Bing AI Performance data?
- Are diagnoses specific to each engine?
- Are documented facts separated from consensus?

## Official documentation
- Perplexity crawlers: https://docs.perplexity.ai/guides/bots
- Bing Webmaster Tools: https://www.bing.com/webmasters
- IndexNow: https://www.indexnow.org/
- Google common crawlers, including Google-Extended: https://developers.google.com/crawling/docs/crawlers-fetchers/overview-google-crawlers
- Microsoft: From Discovery to Influence, a guide to AEO and GEO (Microsoft Advertising, 2026)

---
Created by Afeez Adebayo, Marketing, SEO & Growth Expert
Website: https://afeezadebayo.com
LinkedIn: https://www.linkedin.com/in/afeez-adebayo/
X: https://x.com/za_icon
Source: https://afeezadebayo.com/skills-md/marketing/aeo-geo/answer-engine-platforms/

Licence: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). You may copy, share and adapt this skill, including for commercial use, as long as you credit Afeez Adebayo and link to the source URL above. Keep this notice in any copy or adapted version.

Workflow

Step 1. Access by engine

Check each crawler in robots.txt and logs; verify by IP where providers publish ranges. Remove CDN challenges for search and user-triggered fetchers.

Step 2. Bing foundations

Verify the site in Bing Webmaster Tools (it can import from Search Console), submit sitemaps, turn on IndexNow, fix crawl errors. Check that key pages are indexed in Bing.

Step 3. Read Bing AI Performance

List pages with the most citations and their grounding queries. Note citation share for important queries. Study what cited pages have in common: structure, freshness, data, comparisons.

Step 4. Run the prompts per engine

Record mentions, citations, accuracy and competitors for Perplexity, Copilot, Claude (search on) and Gemini.

Step 5. Diagnose per engine

PatternLikely causeAction
Missing only in CopilotWeak Bing indexing or rankingBing foundations, IndexNow, Bing-specific checks
Missing only in PerplexityPerplexityBot blocked, or no fresh, specific pagesAccess, freshness, specific data
Missing in ClaudeClaude-SearchBot or Claude-User blocked, or weak results for the searches runAccess, answer-ready pages
Missing in Gemini app but present in GoogleGoogle-Extended blockedReview the token decision

Step 6. Improve and re-test

Apply fixes, then re-run the engine's prompts after two to four weeks.

Copilot inside work tools

Many business buyers meet Copilot in Microsoft 365 and Edge, not on bing.com. For B2B brands, Copilot visibility can matter more than consumer traffic data suggests.

Worked example

A B2B HR platform was cited in ChatGPT and Perplexity but never in Copilot. Bing Webmaster Tools showed 40% of key pages not indexed after a migration, and IndexNow wasn't set up. After resubmitting sitemaps, fixing redirect chains Bing still crawled and enabling IndexNow, AI Performance began showing citations for comparison and pricing pages within five weeks, and Copilot named the brand for 7 of 15 tracked prompts.

Output format

# Answer engine review: <brand>, <date>
## Access matrix (engine x crawler: allowed, verified, issues)
## Bing foundations and AI Performance summary
## Prompt results by engine
## Diagnosis and actions per engine, with re-test dates

Testing each engine

EngineTest settingWhat to record
PerplexityDefault search, fresh session, target countryCited sources and their order
CopilotCopilot on web and, for B2B, inside Microsoft 365 or EdgeMentions and linked sources
ClaudeWeb search enabledWhether it searches, sources fetched, description accuracy
Gemini appDefault settingsMentions, links and whether Google sources appear

Engine checklist

EngineMust-have checks
CopilotBing indexing, sitemaps, IndexNow, Bing AI Performance citations
PerplexityPerplexityBot allowed at robots.txt and edge; fresh, specific pages
ClaudeClaude-SearchBot and Claude-User allowed; facts in server HTML
Gemini appGoogle-Extended decision documented; strong Google visibility

Run the checklist quarterly and after any CDN, robots.txt or hosting change, because these engines can lose access silently.

Log every engine change you make, with the date, so later shifts in citations can be explained.

Edge cases and judgement calls

SituationWhat to do
Bing indexes far fewer pages than GoogleFix redirects, sitemaps and quality signals Bing reports; use IndexNow
Perplexity cites forums over your guidesAdd first-hand detail and specifics that forums offer; join the conversation with disclosure
Claude never fetches the siteCheck Claude-SearchBot and Claude-User access at the edge
Gemini app ignores you but Google ranks youCheck whether Google-Extended is blocked and whether that was intended
A partner integration cites you heavily in Bing AI PerformanceStrengthen the cited pages and keep them current
  • Crawler policy: ai-crawler-access
  • Page improvements: answer-ready-content
  • Reporting: ai-visibility-tracking
  • Missing recommendations: ai-citation-earning
  • Bing crawl issues: the Technical SEO redirects-status-codes and xml-sitemaps skills

Quality bar: what an excellent review includes

  • An access matrix per engine and crawler.
  • Bing foundations checked and AI Performance read.
  • Diagnoses specific to each engine.
  • Re-test dates per engine.

Common mistakes

  • Treating all AI engines as one.
  • Ignoring Bing because its search share looks small.
  • Blocking Google-Extended without knowing it affects the Gemini app.
  • Blocking user-triggered fetchers that represent real people.

Questions to ask before you recommend changes

  • Which engines do your buyers use at work and at home?
  • Is AI training access acceptable to the business?
  • Is Bing Webmaster Tools set up and monitored?

Self-check before you answer

  • Did I check access per crawler and verify by IP where possible?
  • Did I use Bing AI Performance data?
  • Are diagnoses specific to each engine?
  • Are documented facts separated from consensus?

Official documentation

When to use this skill, and when not to

Use it when one engine ignores you while others don’t, when B2B buyers meet Copilot inside Microsoft 365 and Edge, when Perplexity referrals matter, and when deciding policies for Google-Extended and Claude’s crawlers.

Each engine works differently. Copilot is built on Bing and gives site owners citation counts by page and grounding query. Perplexity has its own index and cites many sources per answer. Claude searches the web when a question needs it, with separate crawlers for search, user requests and training. The Gemini app grounds with Google, and the Google-Extended token controls whether your content is used there.

Don’t use it for ChatGPT (use ChatGPT Search Visibility) or Google AI Overviews and AI Mode (use AI Overviews and AI Mode).

Example prompts

  • “We appear in Perplexity but never in Copilot. Diagnose it.”
  • “Read our Bing AI Performance export and tell me which pages earn citations and why.”
  • “Should we block Google-Extended? What would we lose?”
  • “Set the robots.txt rules for Claude-SearchBot, Claude-User and ClaudeBot.”
  • “Is IndexNow set up correctly on our WordPress site?”

Example output

The skill returns a fixed structure. A shortened example for an imaginary site:

# Answer engine review: example.com, 10 October 2026

## Access matrix
| Engine | Crawler | Allowed | Verified | Issue |
| Perplexity | PerplexityBot | Yes | Yes | none |
| Copilot | Bingbot | Yes | Yes | 40% of key pages not indexed in Bing |
| Claude | Claude-SearchBot | Yes | n/a | none |
| Gemini app | Google-Extended | Blocked | n/a | limits Gemini app use |

## Bing AI Performance
Top cited: /compare/ (412 citations, 90 days). Pricing pages: 0.

## Actions
Resubmit sitemaps, enable IndexNow, fix redirect chains Bing still crawls.

Diagnoses are specific to each engine, because a fix for Copilot (Bing indexing) does nothing for Perplexity and the reverse.

What the agent needs from you

The skill asks for robots.txt, CDN bot rules and 30 days of logs, Bing Webmaster Tools access and IndexNow status, the prompt set and markets, GA4 referrals from perplexity.ai, copilot.microsoft.com, bing.com, claude.ai and gemini.google.com, and the business’s stance on AI training crawlers.

Research behind this skill

FindingSourceLabel
Perplexity-User generally ignores robots.txt because a user requested the fetchPerplexity crawler documentationConfirmed
Bing AI Performance reports citations, cited pages, grounding queries and citation shareMicrosoft, 2026Confirmed
Google-Extended controls training and grounding in Gemini Apps and Vertex AI, not Google SearchGoogle crawler documentationConfirmed
Copilot visibility follows Bing indexing and rankingPractitioner analysisConsensus

How this skill hands off to related skills

Engine work connects to the core skills. Crawler policy goes to AI Crawler Access, weak pages to Answer-Ready Content, reporting to AI Visibility Tracking and missing recommendations to Earning Citations. Bing indexing issues often come from redirects and sitemaps, covered by Redirects and HTTP Status Codes and XML Sitemaps.

How to run the Perplexity, Copilot, Claude and Gemini skill step by step

  1. Install the skill as answer-engine-platforms/SKILL.md.
  2. Share robots.txt, CDN rules and logs for each engine’s crawlers.
  3. Connect Bing Webmaster Tools, including AI Performance and IndexNow status.
  4. Provide the prompt set and GA4 referrals from each engine.
  5. State your stance on training crawlers such as ClaudeBot and Google-Extended.
  6. Fix per engine and re-test each after two to four weeks.

Situations this skill handles

Each engine fails in its own way. The skill diagnoses these situations:

SituationWhat the skill does
Missing only in CopilotChecks Bing indexing, sitemaps and IndexNow
Missing only in PerplexityChecks PerplexityBot access and freshness of key pages
Claude never fetches the siteChecks Claude-SearchBot and Claude-User at the edge
Gemini app ignores youChecks whether Google-Extended is blocked
B2B buyers use Copilot at workPrioritises Copilot visibility for Microsoft 365 users
Bing AI Performance shows citationsLearns from the cited pages and grounding queries

What makes this skill different

Most AI search advice treats every engine as one. This skill maps each engine’s retrieval source and documented crawlers, then diagnoses and fixes problems engine by engine. It uses Bing Webmaster Tools AI Performance, the only first-party citation data any engine offers, to learn which pages and queries earn citations.

It also explains the token decisions clearly: Google-Extended affects the Gemini app and training, not Google Search, and Claude’s crawlers can be allowed or blocked separately for search, user requests and training.

Practical tips from real projects

  • Bing is a multiplier for Copilot and several partners.
  • IndexNow is quick to set up and helps freshness in participating engines.
  • User-triggered fetchers represent real people; blocking them blocks your customers.
  • Test each engine separately.

A worked example: a B2B HR platform was cited in ChatGPT and Perplexity but never in Copilot. Bing Webmaster Tools showed 40% of key pages not indexed after a migration and no IndexNow. Resubmitting sitemaps, fixing redirect chains Bing still crawled and enabling IndexNow brought citations for comparison and pricing pages within five weeks, and Copilot named the brand for 7 of 15 tracked prompts.

Perplexity tends to favour fresh, specific sources and shows many citations per answer, so updated pages with clear data often appear there first.

Claude searches the web when a question needs current information, and Anthropic documents separate crawlers for search, user-requested fetches and training. The Gemini app grounds answers with Google, so strong Google visibility usually carries over unless Google-Extended is blocked. Writing these decisions into one access policy, engine by engine, stops teams accidentally removing the business from an engine they wanted to appear in.

Frequently asked questions

Allow PerplexityBot, publish specific, current pages that answer the questions your buyers ask and earn mentions on sources Perplexity cites. Freshness matters for time-sensitive topics.

Copilot is built on Bing. Pages that Bing indexes and ranks well for the searches Copilot runs are the ones it cites. Bing Webmaster Tools AI Performance shows which of your pages it cites.

A report in Bing Webmaster Tools showing how often your pages are cited in Copilot, Bing AI summaries and some partner integrations, by page and grounding query, with citation share. It has no click data.

Yes. Anthropic documents ClaudeBot for training, Claude-SearchBot for search and Claude-User for user-requested fetches. Each can be managed separately in robots.txt.

Your content isn’t used for training Gemini models or grounding in the Gemini app and Vertex AI. Google Search, including AI Overviews, isn’t affected.

Yes for Bing and other participating engines. It tells them when pages change, which helps Copilot-related freshness. Google doesn’t use it.

Changelog

  • 1.0.0, 10 October 2026: first release.

Use, adapt and share this skill under CC BY 4.0: credit Afeez Adebayo and link to this page.