geo-competitive-analysis
This skill should be used when the user asks to "analyze competitor AI search visibility", "competitive analysis for GEO", "see who ranks in AI search", "compare AI citations with competitors", "AI search competitive audit", "who does AI recommend over us", "competitive intelligence for generative search", "benchmark against competitors in AI", or any variation of analyzing, benchmarking, or auditing competitor visibility, citations, and brand presence in AI search engines.
GEO Competitive Analysis
GEO competitive analysis answers one question: when a buyer asks an AI engine about your category, who gets mentioned — you or your competitors? Unlike SEO competitive analysis (which compares keyword rankings), GEO competitive analysis compares brand mention rates, citation frequency, and recommendation positioning across AI-generated answers.
Most SaaS companies have never checked who AI engines recommend in their category. By the time they do, competitors may have been cited for months and built entrenched authority. Running a GEO competitive analysis first establishes where you stand and reveals exactly what competitors are doing that you're not.
The GEO Competitive Audit Framework
Step 1: Define the competitor set
Include every company an AI engine might recommend alongside or instead of you.
| Category | Who to include | How many |
|---|---|---|
| Direct competitors | Companies with similar product and same ICP | 3-5 |
| Category adjacents | Companies in adjacent categories AI might confuse with yours | 1-2 |
| Aggregators | Review sites, directories that compete for citations (G2, Capterra) | 2-3 |
Rule: If you're unsure whether a company is a competitor in AI search, test it. Ask ChatGPT "What are alternatives to [your product]?" The answer tells you exactly who AI engines consider your competitive set.
Step 2: Build the query matrix
Create a matrix of queries across intent types:
| Query type | Example queries | Why it matters |
|---|---|---|
| Category definition | "What is [category]?", "How does [category] work?" | Tests who owns the category definition |
| Best-of lists | "Best [category] tools", "Top [category] software 2026" | Tests who gets recommended |
| Comparison | "[You] vs [Competitor]", "[Competitor A] vs [Competitor B]" | Tests how AI positions you vs competitors |
| Alternatives | "[Competitor] alternatives", "Tools like [competitor]" | Tests whether you appear when competitors are mentioned |
| Problem queries | "How to [solve problem you solve]" | Tests whether you're cited in problem-solving context |
| Purchase queries | "[Category] pricing comparison", "How much does [category] cost?" | Tests citation in buying-stage queries |
Target: 30-50 queries covering all intent types and all competitors.
Step 3: Run the audit
Test every query across all three major AI engines.
Tracking template:
| Query | Engine | Your brand mentioned? | Your brand cited (with source link)? | Competitor A mentioned? | Competitor B mentioned? | Who's #1? | Answer accurate? |
|---|---|---|---|---|---|---|---|
| "Best CRM tools" | ChatGPT | Yes (3rd) | No | Yes (1st) | Yes (2nd) | Competitor A | Yes |
| "Best CRM tools" | Perplexity | No | N/A | Yes (1st) | Yes (2nd) | Competitor A | N/A |
Record for each result:
- Mentioned? (brand appears in the generated answer)
- Cited? (AI engine links to a specific page as source)
- Position? (1st, 2nd, 3rd in mention order — order implies recommendation strength)
- Accurate? (facts about your brand are correct)
Step 4: Calculate competitive metrics
| Metric | Formula | What it tells you |
|---|---|---|
| Share of Voice (SOV) | Your mentions / total competitor mentions across all queries | Your relative AI visibility |
| Citation gap | Queries where competitor is cited and you're not | Where you're losing and need to fix |
| Category ownership | Who's cited for "What is [category]?" | Who AI considers the category authority |
| Recommendation rate | % of "best X" queries where you appear in top 3 | How often AI recommends you |
| Comparison win rate | % of "[You] vs [Competitor]" queries where you're positioned favorably | How AI positions you head-to-head |
Diagnosing Why Competitors Win
When a competitor is cited and you're not, diagnose the root cause.
Common competitive gaps
| Gap type | How to identify | Fix |
|---|---|---|
| Content gap | Competitor has a page for the query topic, you don't | Publish a page targeting that query |
| Structure gap | Both have pages, but competitor's is more extractable | Apply AEO formatting: answer-first, tables, Q&A, schema |
| Recency gap | Your page is older than competitor's | Update your page with fresh data, new dateModified |
| Authority gap | Competitor has more third-party mentions, reviews, press | Invest in GEO source mentions — reviews, guest posts, podcast appearances |
| Entity gap | AI knows your competitor better than it knows you | Build entity signals: Wikidata, Organization schema, consistent branding |
| Schema gap | Competitor has structured data, you don't | Add FAQPage, Article, and relevant schemas to your pages |
Competitor page analysis
When a competitor is cited for a specific query, read their page and score it:
| Dimension | What to check |
|---|---|
| First 50 words | Do they answer the query immediately? |
| H2 structure | Are H2s question-shaped and matching the target query? |
| Tables | Do they use tables for comparisons and data? |
| Schema markup | Check page source for JSON-LD schemas |
| Author + date | Named author? Published and modified dates? |
| Content depth | Word count, original data, unique insights |
| Tone | Balanced (covers competitors honestly) or one-sided? |
The fix is usually structural, not topical. Most citation gaps come from format and structure differences, not from one company knowing more than another.
Competitive Monitoring
Weekly monitoring (15 min)
- Spot-check 5 priority queries across engines
- Flag any new competitors appearing in your category queries
- Note any changes in your mention position
Monthly competitive report
| Section | Content |
|---|---|
| SOV trend | Your share of voice vs top 3 competitors, trended over time |
| Gains | Queries where you gained a citation this month |
| Losses | Queries where you lost a citation this month |
| New threats | Competitors who appeared for the first time |
| Action items | Top 5 fixable gaps for next month |
Quarterly deep audit
Full re-run of the GEO competitive audit:
- Re-test all 30-50 queries across all engines
- Recalculate all competitive metrics
- Re-analyze top competitor pages for structural changes
- Update strategy based on competitive shifts
- Present findings to stakeholders
Competitive Response Playbook
When a competitor is cited and you're not
| Urgency | Situation | Response |
|---|---|---|
| High | Competitor cited for "[You] vs [Competitor]" — your own comparison query | Publish or rewrite your comparison page within 1 week. AEO-optimize with answer-first structure, tables, schema |
| High | Competitor cited for "Best [category] tools" — category leadership query | Publish comprehensive listicle/comparison page. Build review volume. Add category definition content |
| Medium | Competitor cited for "[Competitor] alternatives" — you're not listed | Publish your own "[Competitor] alternatives" page with your brand included |
| Medium | Competitor cited for problem queries you should own | Publish targeted how-to content answering the specific problem query |
| Low | Competitor cited for their own brand queries | Normal. Focus on your own brand queries instead |
When a competitor's AI description is wrong about you
AI engines sometimes describe your product inaccurately in comparison to competitors. This is fixable.
Process:
- Document the inaccuracy (screenshot, exact text)
- Identify the likely source (which page is the AI reading?)
- If the source is your own page — fix the content to make the correct fact extractable
- If the source is a competitor's page — publish your own page with the correct information, better structure, and more authority signals
- If the source is a third-party site — reach out to correct the information, or publish your own authoritative page that outcompetes it
- Re-test in 2-4 weeks
Pre-Audit Checklist
Before running a GEO competitive analysis:
- [ ] Competitor set defined (3-5 direct + 1-2 adjacent + 2-3 aggregators)
- [ ] Query matrix built (30-50 queries across all intent types)
- [ ] Testing accounts set up for ChatGPT, Perplexity, and Gemini
- [ ] Tracking spreadsheet created with all required fields
- [ ] Baseline measurement plan defined (test over 2-3 sessions for reliability)
- [ ] Responsible person assigned for ongoing monitoring
- [ ] Response playbook shared with content team
- [ ] Reporting cadence set (weekly spot-check, monthly report, quarterly deep audit)
- [ ] Access to competitor pages for structural analysis
Anti-Pattern Check
- Never checking who AI recommends in your category → You might be invisible while competitors get recommended daily. Run the basic 30-query audit this week. It takes 2-3 hours and shows you exactly where you stand
- Only testing in one AI engine → ChatGPT, Perplexity, and Gemini cite different sources and recommend different products. A competitor winning in Perplexity but losing in ChatGPT requires different tactics than one winning everywhere
- Assuming your SEO position = your AI position → A company ranking #1 in Google for your category keyword may not be the one AI engines recommend. GEO and SEO have different authority signals. Always test AI engines directly
- Reacting to individual test results instead of trends → AI engine responses vary between sessions. Don't panic over one bad result or celebrate one good one. Track trends over weeks and months
- Copying competitor content instead of improving on it → If a competitor is cited because they have a great comparison table, don't copy their table. Build a better one — more complete, more accurate, more recent, with original data they don't have
- Only monitoring, never acting → A competitive report that sits in a Google Doc helps no one. Every monthly report should produce 3-5 specific action items with owners and deadlines
Want agents that use skill files like this?
We customize skill files for your brand voice and methodology, then run content agents against them.
Book a call