AI Search (AEO/GEO): Get Found by Machines That Answer
20 slides on why search traffic is collapsing, where it’s going, the Signal + Content framework for AI visibility, how to track it — and where this ends.
20 lessons · ~4 min · free, no signup.
Lessons
- AEO / GEO — DEFINITION
- The prediction that came true — STAT
- Traffic is changing owners — STAT
- Zero-click searches — STAT
- How this differs from SEO — DEFINITION
- Small traffic, huge intent — STAT
- The two levers: Signal + Content — DEFINITION
- Citations come from earned media — STAT
- Content that gets cited — PLAYBOOK
- llms.txt — DEFINITION
- Know your bots — DEFINITION
- The measurement blind spot — STAT
- Only 16% track it — STAT
- How to track AI visibility — PLAYBOOK
- Queries got longer — STAT
- B2B has already moved — STAT
- MCP as marketing infrastructure — DEFINITION
- Myth: SEO is dead — MYTH
- The 30-day AEO sprint — PLAYBOOK
- You can now see the invisible traffic. — COMPLETE
Full course text
1. AEO / GEO
Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO): making your brand the source AI systems cite when they answer questions — in ChatGPT, Claude, Perplexity, Gemini, and Google’s AI Overviews. SEO earned rankings. AEO earns citations.
Source: Robauto — AI Visibility Platform
2. The prediction that came true
In 2024 Gartner predicted traditional search volume would drop 25% by 2026. A 2026 study of 1M+ keywords found 29% of search demand already in decline — worse than forecast.
Source: Gartner (2024); Fractl × Search Engine Land (2026)
3. Traffic is changing owners
Google referral traffic to publishers fell ~33% globally in 2025 (−38% in the U.S.) while AI referral sessions grew ~1,200% in the same period. The traffic isn’t disappearing — it’s changing owners.
Source: Chartbeat via Press Gazette (2025)
4. Zero-click searches
58.5% of U.S. Google searches now end without a single click to any website. When an AI Overview appears, organic click-through rates drop ~61%. Ranking #1 under an answer nobody scrolls past is the new page two.
Source: SparkToro / Similarweb; BrightEdge CTR data
5. How this differs from SEO
SEO optimizes for a ranked list of links and wins clicks. AEO optimizes for a synthesized answer and wins mentions, citations, and recommendations. Different crawlers, different signals — and only ~11–12% overlap between Google’s top results and what ChatGPT cites.
Source: Ahrefs / QuickSEO citation-overlap studies (2026)
6. Small traffic, huge intent
AI referrals are only ~1% of web traffic today — but they convert dramatically better. Adobe found AI-referred retail visitors converted 42% better than non-AI traffic by March 2026 (a year earlier they converted 38% worse).
Source: Adobe Digital Insights (2026)
7. The two levers: Signal + Content
Robauto’s framework. SIGNAL = everything machines read about you off-page and on-page: structured data, third-party mentions, llms.txt, feeds, MCP endpoints, crawlability. CONTENT = what’s worth citing: original stats, depth, answers, freshness. Signal gets you read; Content gets you quoted.
Source: Robauto — Five-Pillar Visibility Model
8. Citations come from earned media
About 84% of AI citations point to earned, third-party media — not brand-owned pages. Brands mentioned across publishers, reviews, and communities earn ~6.5× more AI citations than brands who only publish on their own domain. PR is now a technical SEO channel.
Source: Muck Rack (2026); University of Toronto (arXiv:2509.08919)
9. Content that gets cited
Original statistics (+156% citation probability), quotes and cited sources (+30–40% visibility), structured Q&A blocks (+44%), author schema (3×), and freshness — pages updated within 60 days are ~1.9× more likely to be cited. Publish data, not adjectives.
Source: BrightEdge / Authoritas / Superlines (2025–26)
10. llms.txt
A plain-text file at yourdomain.com/llms.txt that gives AI crawlers a curated map of your most important content — like robots.txt’s welcoming cousin. Pair it with llms-full.txt (full content dump) and a robots.txt that explicitly allows AI bots.
Source: llmstxt.org proposal
11. Know your bots
GPTBot (OpenAI), ClaudeBot (Anthropic), PerplexityBot, Google-Extended, CCBot (Common Crawl). Blocking them = invisibility. Most analytics can’t even see them — which is why bot-level detection (like Robauto’s tracking pixel) is the first instrument to install.
Source: OpenAI / Anthropic crawler docs
12. The measurement blind spot
35–70% of AI-referred sessions arrive with no referrer data and get miscounted as ‘Direct’ traffic. GA4 only added a native ‘AI Assistant’ channel in May 2026 — and it still misses several engines. Whatever AI traffic you’re reporting, it’s an undercount.
Source: Adwen Plus practitioner data; GA4 changelog (2026)
13. Only 16% track it
Just 16% of Fortune 500 companies systematically track their AI search performance. For small and mid-size businesses, it’s far lower. The scoreboard exists; almost nobody is watching it. That’s the opportunity.
Source: AirOps / McKinsey (2026)
14. How to track AI visibility
1) Detect AI crawler hits on your site (pixel/log analysis). 2) Segment AI referrals in analytics (chatgpt.com, perplexity.ai, claude.ai). 3) Prompt-test: ask each engine your money questions monthly and log whether you’re cited. 4) Score it: Robauto’s five pillars.
Source: Robauto — Visibility Scoring
15. Queries got longer
AI Mode queries run ~3× longer than traditional searches. ‘Best running shoes 2026’ becomes ‘I’m training for my first 5K and have never bought running shoes — where do I start?’ Your content must answer situations, not keywords.
Source: Google AI Mode usage report (2026)
16. B2B has already moved
94% of B2B buyers used generative AI during their purchase process, and roughly half now start vendor research in an AI chatbot instead of Google. By the time they reach your site, the shortlist already exists.
Source: 6sense Buyer Experience Report (2025); G2 (2026)
17. MCP as marketing infrastructure
The frontier of AEO: don’t just be readable by AI — be usable. A white-label MCP server turns your site into a live tool AI assistants can query directly (inventory, guides, calculators). The brands machines can interact with beat the brands machines can only read about.
Source: Model Context Protocol · Robauto white-label MCP
18. Myth: SEO is dead
No — it’s demoted from strategy to substrate. ~76% of Google AI Overview citations still rank in the top 10, and 21.6% of ChatGPT’s outbound clicks go back to Google. The journey is now AI → Google → you. Do both; measure separately.
Source: Semrush AI Overview studies; Similarweb (2026)
19. The 30-day AEO sprint
Week 1: allow AI bots, ship llms.txt + schema, install bot detection. Week 2: rewrite top 10 pages as direct answers with original data and dated updates. Week 3: earn 3 third-party mentions. Week 4: prompt-test 20 buyer questions across 4 engines; baseline your score.
Source: Robauto AEO methodology
20. You can now see the invisible traffic.
Search didn’t die — it split. You know the numbers, the Signal + Content levers, and the tracking stack. Next: where AI leaves the screen entirely.
Source: Continue → IoT & Robotics