Heplink
    Back to insights
    Strategy

    AI SEO in 2026: how to show up in ChatGPT, Perplexity & Google's AI Overviews

    20 June 2026
    11 min read

    Search isn't ten blue links anymore. People ask ChatGPT, Perplexity, Gemini and Google's AI Overviews — and a single cited brand often wins the whole answer. Here's the playbook we use to get clients pulled into those answers.

    The shift in plain English

    Classic SEO was about ranking a page on a results list. AI SEO — sometimes called Generative Engine Optimisation (GEO) — is about being cited inside the answer an LLM gives the user. There's no "page 2". You're either in the answer or invisible.

    The good news: the underlying signals overlap with traditional SEO (authority, clarity, structured content, fresh data). The new part is making your content easy for an LLM to read, quote and attribute.

    Where the answers come from

    ChatGPT (with browsing & SearchGPT)

    Pulls from Bing index + cited live sources. Citations show up as inline links.

    Google AI Overviews & Gemini

    Pulls from Google's standard index, biased toward high-authority and structured pages.

    Perplexity

    Lives or dies on citations. Almost every answer links sources, so being cited = traffic.

    Claude, Copilot, Meta AI

    Mostly retrieval-augmented — they fetch and quote, then attribute (or paraphrase) sources.

    The 7-part AI SEO playbook we use

    1. Build content around real questions, not keywordsLLMs are asked questions in full sentences. Write pages that directly answer them with the answer near the top, then expand. Use H2/H3 phrased as the question.
    2. Front-load the answer (TL;DR-first writing)Models extract the most concise correct paragraph. Open every section with a 2-3 sentence direct answer before you elaborate. This is how you become the quoted line.
    3. Add original data, opinions and named entitiesLLMs over-cite content that contains things they can't generate themselves: original stats, named experts, specific case studies, dated numbers, real quotes.
    4. Ship structured data and clean semantic HTMLArticle, FAQ, HowTo, Organization and Person schema. Real <h1>, <h2>, <ul>, <table>. Models parse the DOM — messy markup gets skipped.
    5. Publish an llms.txt (and keep it up to date)An /llms.txt at your root tells AI crawlers what pages matter and how to describe you. It's the robots.txt of the LLM era. Heplink has one — most sites still don't.
    6. Earn citations from sites LLMs already trustWikipedia, Reddit, G2, industry publications, news. Models heavily weight these as sources. Digital PR and genuine community presence is now an AI SEO tactic.
    7. Monitor visibility in AI answersTools like Profound, Otterly, AthenaHQ and manual prompt testing let you track which models are citing you for which prompts. Treat it like rank tracking — but for answers.

    The mistakes killing brands in AI search

    • • Writing for keyword density instead of clarity. LLMs reward the cleanest answer, not the densest page.
    • • Hiding key facts behind tabs, accordions or JS-only renders. If it isn't in the initial HTML, expect it to be missed.
    • • No author bios, no organisation schema, no E-E-A-T signals. Models heavily favour sources they can identify.
    • • Treating AI Overviews as "just another SERP feature". When it appears, average organic CTR drops sharply — you have to be inside the Overview.
    • • Ignoring Reddit, YouTube transcripts and forums. These are over-represented in LLM training and retrieval.

    What "winning" looks like in 2026

    For one of our SaaS clients, being cited by Perplexity and ChatGPT now drives more qualified demo requests than classic Google organic. Volume is lower — but intent is miles higher. People arrive having already read an answer that recommended them.

    That's the real shift: AI SEO isn't another channel. It's the new top of funnel for considered purchases. Brands that treat it seriously now will look untouchable in 18 months. Brands that don't will quietly disappear from the answer.

    Quick checklist

    • ✓ Publish /llms.txt and /sitemap.xml.
    • ✓ Add Article, FAQ and Organization schema to every meaningful page.
    • ✓ Rewrite top 10 commercial pages with TL;DR-first answers.
    • ✓ Add original data, dates and named authors.
    • ✓ Track 25 priority prompts in ChatGPT, Perplexity and Google AI Overviews monthly.

    The brands that win AI search in 2026 are the ones writing for the model and the human in the same sentence.

    Want to be cited by ChatGPT and Perplexity?

    We run AI SEO audits and implementation for ambitious UK brands — schema, content rebuilds, llms.txt, citation strategy and monthly visibility tracking.

    Book a strategy call