AI Search SEO: AEO, GEO and the new generative search
How to optimise your site to appear as a cited source in ChatGPT, Perplexity, Gemini and Google AI Overviews. Operational method with 15 technical and editorial checkpoints.
AEO (Answer Engine Optimization) and GEO (Generative Engine Optimization) are the 2 disciplines replacing traditional SEO in 2026. LLMs prioritise: (1) content with extractable TL;DR, (2) hard data with source, (3) dense schema.org (Article + FAQPage + HowTo + Speakable), (4) author authority (Person schema + LinkedIn), (5) freshness (recent updatedAt). Not appearing in LLMs means losing 30-45% of informational traffic in 2026.
What are AEO and GEO?
AEO (Answer Engine Optimization) is the discipline optimising content to be cited as an answer by direct-answer engines: Google AI Overviews, Bing Copilot, Perplexity, You.com. The goal is to appear in the answer box, not just among the 10 blue links.
GEO (Generative Engine Optimization) is its extension for conversational LLMs (ChatGPT, Claude, Gemini): appearing as a cited source when the user asks a question in chat. It differs from classic SEO because LLMs don’t crawl in real time: they consume aggregated indexes and prioritise authority signals, not backlinks.
AEO + GEO method in 15 checkpoints
The 15 checkpoints are grouped in 4 blocks: content, structure, schema and authority. Apply them by priority: blocks 1 and 2 deliver 70% of the result.
- TL;DR of ≤ 55 words at the startLLMs extract the first 55-80 words as summary. Write a canonical TL;DR that answers the main question in 2-3 sentences with at least 1 hard number.
- Canonical answer blockAfter the TL;DR, a "What is X?" block with formal definition in 1-2 paragraphs. LLMs literally quote this block when the user asks the definition.
- Hard data with citable sourceEvery numeric datum must carry context and (if possible) source. "CPL 47% below sector average" is citable; "very low CPL" is not. LLMs prioritise concrete data.
- Headings with explicit questionsUsing H2/H3 as questions ("How much does an insurance lead cost?") improves citation 30-45% vs affirmative headings. LLMs map question→answer.
- Structured FAQ at the end6-10 FAQs at article end, each with a direct answer. LLMs use them as long-tail citation targets.
- Complete schema.org ArticleArticle with headline, author, publisher, datePublished, dateModified, wordCount, timeRequired, articleSection and image. Without these fields the post is not indexed by AEO systems.
- schema.org FAQPageDuplicate the visible FAQ in schema:FAQPage. Google AI Overviews cites FAQPage more frequently than plain text.
- schema.org HowTo when applicableIf the article describes a step-by-step method, emit HowTo with numbered HowToSteps. Perplexity and Gemini prefer HowTo for "how to X" queries.
- schema.org SpeakableSpecificationMark TL;DR and definition as speakable. Voice assistants (Alexa, Google Assistant, Siri) preferentially read speakable blocks.
- Person schema for authorsEach author with complete Person (name, jobTitle, sameAs with LinkedIn, worksFor). Direct E-E-A-T signals that LLMs value for citation.
- Content freshness (updatedAt)Update dateModified with real changes every 60-90 days. LLMs penalise stale content. Adding an "Update YYYY-MM" section with change notes helps.
- Internal links with descriptive anchorEach guide links to 3-5 site sections with anchor describing the destination page. LLMs follow links to expand context.
- Speed and CLS optimisationLCP < 1.5s, CLS < 0.05, INP < 200ms. LLMs consume Chrome UX Report metrics; slow sites are dropped from AEO indexes.
- llms.txt at the domain rootFile /llms.txt with priority pages for LLMs (equivalent to sitemap.xml but LLM-oriented). Adopted by ChatGPT, Perplexity and Anthropic since 2024.
- Presence in aggregate indexesAppear in Wikipedia (references), Wikidata (entity), sector directories with schema.org. LLMs cross-reference entities between indexes.
Frequently asked questions
Is AEO the same as SEO?
No. SEO optimises for organic position in blue results; AEO optimises for appearance in the direct answer box. They share fundamentals (quality content, schema, speed) but differ in specific techniques (TL;DR, answer block, speakable, llms.txt).
How long does it take to see AEO results?
Perplexity and ChatGPT update indexes every 4-8 weeks. First citation results appear between week 6 and 12 after optimised publication. Google AI Overviews takes 3-6 months.
Can you measure traffic from ChatGPT and Perplexity?
Yes, with UTM parameters and user-agent filters in GA4. ChatGPT sends user-agent "ChatGPT-User" and Perplexity "PerplexityBot". You can create custom channels in GA4 to measure LLM-sourced sessions.
What is llms.txt and why should I have one?
Standard file proposed by Answer.AI (2024) at /llms.txt of the domain root, with curated list of priority pages for LLMs. Adopted by ChatGPT, Perplexity, Anthropic. Its absence does not penalise but its presence accelerates LLM indexing.
Do AEO techniques change every quarter?
Yes, at the same frequency as Google updates its algorithm (2-4 core updates/year) but with less public documentation. Serious operators audit their AEO content every 90 days.
How does AEO affect classic SEO?
Positively. AEO checkpoints (dense schema, FAQ, HowTo, Speakable, author with Person) also reinforce classic Google E-E-A-T signals. Content well optimised for AEO tends to rise 15-30% in organic positions.
How to appear as a cited source in ChatGPT, Perplexity, Gemini and Google AI Overviews in 2026
Search traffic has begun moving from the 10 blue links toward direct answer boxes and LLM chats. 2025 studies indicate Google AI Overviews appear in 47% of Spanish informational queries, and ChatGPT + Perplexity together handle over 400M daily queries with potential to cite external sources.
Sites cited in LLMs multiply their "post-AI" traffic by 2-4×, while sites optimised only for classic SEO lose between 30% and 45% of informational traffic. At eXprimeNet we have operated AEO + GEO since 2023 and audit digital-acquisition sites to position them as citable sources.