Mike Wilhelm

Generative Engine Optimization

3 min read

Sources: Ishaan Shakunt (SpearGrowth.com, Cosenly.com), Andrei Țiț, (Head of Product Marketing at Ahrefs)

Strategy: What AI Search Actually Rewards

AI Search ≠ Traditional SEO

  • Less than 30% overlap between Google SEO and ChatGPT recommendations
  • SEO is about popularity. AI search is about factual relevance and semantic fit
  • Traditional SEO rewards repetition and precision
  • Generative engines prioritize clarity, structured formatting, and content density

Takeaway: What works for SEO often doesn’t work for Generative Engine Optimization (GEO)

What Not to Do (Baseless Busywork)

  • Fixing schema markup
  • Adding FAQs
  • Uploading LLMs.txt
  • Making content more conversational
  • Building backlinks

These actions have little impact on AI visibility.

Target Only BOFU Searches

Forget TOFU and MOFU content. Only optimize for purchase intent queries:

  1. Recommend software like this
  2. Help me choose between A & B
  3. Tell me about company C
  4. Top 5 [category] tools

Examples:

I’m a CFO looking for a financial reporting tool that integrates with ERP X.
I’m choosing between BambooHR and Zoho People. Which has better compliance?
I run demand gen and want to know if Chosenly.com improves AI search performance.

Create Criteria, Not Just Keywords

You can’t predict AI search queries. Instead, map out what criteria LLMs use to recommend:

  1. Ask your sales team what prospects might ask an LLM
  2. Discover what LLMs currently believe about your category
  3. Map where you rank well or fall short in their perceived criteria
  4. Identify areas where you have a right to win
  5. Add supporting facts to your site and third-party pages
  6. Monitor rankings monthly using variations and dummy accounts

If AI gets facts wrong about you, find and fix the source — even if it’s on someone else’s site.

Content: What and How to Publish

Build Fact-Dense, Structured Content

  • Use bullets, summaries, headers, and “in summary” phrasing
  • Focus on clarity, high sentence count, and semantic depth
  • Don’t keyword-stuff
  • Create content that aligns closely with AI prompts

Avoid Copycat Content

  • Don’t write generic “What is…” articles
  • Informational content is dying in this context
  • Focus on differentiated BOFU topics that answer real, intent-driven questions

Visibility: How to Get Cited by LLMs

Approach 1: Be Where LLMs Pull From

  • Top 10 listicles
  • Review sites (G2, Capterra, SelectHub)
  • Wikipedia profiles
  • PR/media coverage
  • Market research (e.g., Financial Times)

ChatGPT does look at Reddit and LinkedIn — but rarely for B2B queries.

Approach 2: Replace Weak Citations

  • Perform targeted searches
  • Document vague or incorrect citations in a spreadsheet
  • Create or update content to replace/improve those sources

Tactics for Visibility

  • Don’t block LLM crawlers (they often ignore llm.txt and robots.txt)
  • Publish deep, comprehensive, semantically aligned content
  • Focus on:
    • Word count
    • Sentence count
    • Domain rating
  • Use brand mentions with rich anchor text
  • Be active on crawlable communities (Reddit, Quora)
  • Consider a PR strategy resurgence for AI visibility

Measurement: How to Track AI Search Impact

Manual Monitoring

Track monthly:

  • If you’re being cited
  • Where citations come from
  • Hallucinations or incorrect info

Google Analytics 4 Setup for ChatGPT

  1. Go to Reports > Lifecycle > Traffic Acquisition
  2. Change dropdown to Session source
  3. Search for chatgpt
  4. (Optional) Save as a custom segment
  5. Add Landing page + query string as a secondary dimension to see which pages ChatGPT links to

Performance Benchmarks

  • AI search sends about 0.1% of total referral traffic
  • Don’t expect volume — expect brand awareness and qualified traffic
  • For Ahrefs:
    • 0.5% of traffic comes from AI search
    • 12% of signups come from it

Strong correlation: Brand search volume boosts AI search visibility.