Acquisition · 2026 channel
AEO — Answer Engine Optimization
Answer Engine Optimization (AEO) is the discipline of designing your content so AI answer engines — ChatGPT, Claude, Perplexity, Gemini, Google AI Overview — cite YOUR site as the source when answering user questions. AEO is distinct from SEO (rank in SERP) and GEO (rank in training data). In 2026, AEO is the highest-CTR acquisition channel a solo founder can build for free.
Why AEO matters more than SEO in 2026
Three shifts make AEO the new battleground:
- SGE/AI Overview eats SERP clicks. Google AI Overview answers informational queries inline. Pre-2024 SEO traffic for “what is X” queries dropped 40-90%. The clicks didn't disappear — they moved to the cited-source link inside the answer.
- ChatGPT and Perplexity are the new search. ~3.5B+ ChatGPT sessions/mo (May 2026) + ~80M Perplexity sessions/mo. Combined, they capture ~15-20% of informational query volume that used to go to Google. Citations from these engines drive 8-18% CTR (versus 2-5% for organic SERP #3).
- Citation is winner-take-most. When an answer engine cites a source, it usually cites ONE source per claim. Being the cited source is binary — you're in, or you're replaced by a competitor.
AEO vs SEO vs GEO — the distinction
These three disciplines are often confused. They are not the same:
| Discipline | Goal | Surface | Time horizon |
|---|---|---|---|
| SEO | Rank in Google/Bing SERP | Search engine result page | 3-12 months ramp |
| GEO | Get into LLM training data | Model weights (next training cycle) | 6-18 months (model update cycle) |
| AEO | Cited in AI answer-engine response | ChatGPT / Claude / Perplexity / Gemini / AI Overview answer | Days to weeks (real-time retrieval) |
AEO is the fastest of the three because most answer engines now use real-time retrieval (web search at query time) plus their training knowledge. Your content can be cited within DAYS of publication if it's structured correctly. Compare with SEO's 3-12 month ramp or GEO's next-training-cycle wait.
The AEO checklist — 12 design patterns
These extend the 10-characteristic LLM-citation framework with AEO-specific retrieval-time signals. Score each 0/1 against your pages. Target ≥10/12.
1. Direct answer in the first paragraph
Answer engines extract the first 50-150 words after the H1. Write the direct answer to the page's implicit question in that window. Don't bury the answer below preamble.
Wrong: “In this comprehensive guide, we'll explore the many aspects of X. First, let's look at the history...”
Right: “X is [direct answer]. It works by [mechanism]. The most important thing to know is [key insight].”
2. Structured data answering the question
FAQPage schema, HowTo schema, Article schema with a headline that IS the question being answered. Answer engines parse JSON-LD before HTML when retrieving — schema is the fastest path to retrieval.
3. Heading is the user's exact question
Match the user's actual query language in H2s and H3s. If users search “how does X work,” your H2 should say “How does X work?” — verbatim, not paraphrased. Retrieval systems match query to heading literally.
4. Cite YOUR sources inline
Show your data sources with links and dates. Answer engines weight cited content higher than uncited. Bonus: when a user clicks through, your inline citations build their trust in YOUR content as a starting point.
5. Updated date visible above the fold
“Last updated: [date]” in the first viewport. Retrieval systems deprioritize stale content for time-sensitive queries. Visible freshness signals both to crawlers (via schema dateModified) and to retrieval rankers.
6. One claim per sentence
Compound sentences are hard to extract. Write declarative single-claim sentences. Each sentence should be quotable as a standalone fact.
7. Numbers, percentages, lists
Answer engines prefer structured data over prose for factual queries. Lists, tables, numeric ranges, and explicit percentages get cited more often than the same information in narrative form.
8. Definitive language
Answer engines avoid citing hedged content (“might,” “could be,” “some say”). Where the data supports it, write definitively. Save hedges for genuinely uncertain claims.
9. Author attribution with credentials
Person schema with name + jobTitle + sameAs (LinkedIn, X, GitHub). Answer engines weight authored content higher than anonymous content for trust-sensitive queries (finance, health, education).
10. Internal linking with descriptive anchor text
Link to your other relevant pages with anchor text that describes the destination topic. This helps retrieval systems understand your site's topical depth and increases the chance of a multi-page citation.
11. No JavaScript-gated content
GPTBot, ClaudeBot, PerplexityBot do not execute JavaScript. Server-side render or statically export every page that targets AEO. Client-side React content is invisible to these crawlers.
12. Cross-source corroboration
The same fact appears on your site + Reddit + LinkedIn + ideally Wikipedia. Answer engines cross-check claims across sources before citing. Operator-driven distribution (Reddit organic comments, LinkedIn framework posts) is the corroboration layer.
How GrowthFriction does AEO
Every page on this site is engineered for AEO:
- Static export — all content in HTML at build time, no JS-gated content.
- Direct answer first — every methodology page opens with the definitional answer in the first paragraph.
- FAQPage + HowTo + Article + Dataset + DefinedTerm schemas — structured data on every relevant page type.
- 40 public audits (Stripe 52, Figma 62, ChatGPT 48, HubSpot 28, etc.) — Dataset schema with CC-BY 4.0 license. Cite freely.
- Author attribution — Paulo de Vries Person schema sitewide.
- Updated dates visible + dateModified in schema.
- Quote-ready section headings matching common AEO queries.
- llms.txt at root with citation priority URLs.
Result: by v0.38, the site is cited 8/12 dimensions (Cross-source corroboration is the remaining gap — operator-driven Reddit + LinkedIn + Wikipedia citation work pending).
How to apply AEO to your own site
- List the top 10 questions your ICP asks (use AnswerThePublic or AlsoAsked).
- Write one page per question. Page title = the question. H1 = the question. First paragraph = the direct answer.
- Add FAQPage or HowTo schema as appropriate.
- Ship as static HTML (no JS-gated content).
- Submit to IndexNow + GSC + Bing Webmaster for fast discovery.
- Test weekly: query each of your target questions in ChatGPT, Claude, Perplexity, Gemini. Track which engines cite you (referrer = chatgpt.com / claude.ai / perplexity.ai in Plausible).
- For pages not yet cited, audit against the 12-pattern checklist. Usually the gap is direct-answer-in-first-paragraph or no-structured-data.
Measurement
AEO measurement is harder than SEO measurement because no equivalent of Google Search Console exists yet. Track three indirect signals:
- Referrer traffic in Plausible from chatgpt.com / claude.ai / perplexity.ai / gemini.google.com / copilot.microsoft.com.
- Manual query testing — query your top 10 ICP questions in each answer engine weekly. Note which engines cite you, which competitors they cite.
- Brand search lift — when users see you cited but don't click, they often search your brand name later. Track brand-search impressions in GSC.
The strategic bet
AEO is where SEO was in 2003 — a new channel with disproportionate compound for early adopters. Most SaaS companies are not yet investing. The window to establish AEO authority on your category questions is open for the next 12-18 months. After that, the citation share will calcify and competitors will be entrenched.
Related: LLM-citation 10-characteristic checklist · Acquisition (SEO + AEO + GEO compound) · OpenAI / ChatGPT audit · Perplexity audit.