GEO Checklist

GEO stands for Generative Engine Optimisation — the practice of getting your content cited inside AI answers from ChatGPT, Claude, Perplexity, Gemini, Copilot, and Google's AI Overviews. The mechanics overlap with classical SEO (technical foundations, schema, content quality) but the targets are different: SEO ranks pages, GEO earns the quote inside an AI's synthesised answer.

Pick your industry from the picker (top right of the toolbar) and every check gets reweighted by what actually correlates with ChatGPT recommendations in your vertical. Items that track the industry's strongest signal get a Critical badge and float to the top of their tier. Items tracking a signal that's *negative* in your vertical (e.g. Wikidata in senior home-care, Reddit in pest control) get a Counter-productive badge and drop to the bottom — so you don't waste budget on tactics that hurt you. The data is from the OppAlerts LLM Ranking Factors study (Spearman correlation across 13 web signals × 144 industries × ChatGPT 5.4, March 2026). The reweight only changes priority, never hides the universal items — Tier I writing rules and Tier III technical hygiene apply to everyone.

OppAlerts covers 144 industries — the most-asked-about commercial verticals across SaaS, eCommerce, services, brands, insurance, healthcare, hospitality, and trades. If your exact niche isn't there, pick the closest analogue (e.g. Amazon-seller services has no direct match, but eCommerce SaaS or DTC brand verticals share most of the citation dynamics). When no industry is selected, the universal default applies and every check has equal weight — that's the baseline we built first, and it stays valid.

No — and you shouldn't try in one quarter. Tier 3 (Crawlability) is non-negotiable: if AI crawlers can't reach you or parse your HTML, nothing else matters. Start there, then Tier 1 (the writing rules), then Tier 4 (so you can measure progress), then Tier 2 (the slowest but most compounding tier). Run Tier 5 (Industry calibration) before any of the others to know which signals matter most for your vertical — the lookup takes 10 minutes and reweights everything downstream. The realistic order is Tier 5 lookup in week one, Tier 3 fixes in week one, Tier 1 across a quarter, Tier 4 by month two, Tier 2 over the year.

No account. Your checked items are stored in your browser's local storage on your device. Reload the page, close the tab, come back next week — it's still there. Switch to a different browser or device and it's gone, because it never leaves your machine. We never see your progress and don't track it.

It's a clean, branded PDF that captures your current progress — checkboxes filled in based on what you've completed. Useful for circulating inside a team, attaching to a quarterly review, printing and pinning above a desk, or handing to a contractor as a brief. The print view strips out the interactive UI and lays the content out for paper, with each tier starting on a new page.

Quarterly at minimum. During an active growth phase — building out AI visibility from zero, migrating a site, or recovering from a crawlability regression — monthly. Tier 3 items (Crawlability) need a monthly pass regardless, because robots.txt edits, Cloudflare setting changes, and new bot launches can break coverage without warning. Tier 2 items (Trust signals) compound over 6–12 months, so you're checking progress, not re-working them each quarter.

Some. Several Tier 3 and Tier 4 items are scriptable: FCP (T3.4) via the PageSpeed API, crawler access (T3.1, T3.2, T3.11) via curl + user-agent sweeps, server-log analysis (T4.2) via GoAccess or a Vector/Loki stack, GA4 referrer regex (T4.1, T4.11) via the GA4 API. Our own tool at haide.digital/tools/signal-check automates six of the signals here — crawler access, answer capsule, structured data, First Contentful Paint, technical integrity, and entity clarity. Tier 1 writing quality and Tier 2 trust signals require human judgement. Don't trust anyone who claims to fully automate either.

This page updates quarterly, so the bot list (T3.11), referrer regex (T4.11), and platform-specific items (T4.7, T2.16) stay current. The core principles don't change with new platforms — entity signals, crawlability, freshness, and extractable passages transfer directly. A new LLM from a new provider in Q3 2026 will almost certainly need the same five things the April 2026 roster needs: robots.txt access, a grounding source to search (Brave, Bing, or a proprietary index), Article/Organization schema to parse, visible dates, and passages it can lift without surrounding context.

AI-search behaviour shifts every few months. We re-run the source studies (Indig, RESONEO, AirOps × Indig 2026 State of AI Search, Cloudflare, our own server-log analyses) and refresh items where the underlying data has changed. The page lists the last update date in the changelog at the bottom. If something feels stale or you've spotted a new pattern, message Adrian on LinkedIn — that's how most updates start.

Yes. The checklist is written for founders, product owners, marketing leads, and SEO specialists alike. Every technical term is introduced before it's used. You don't need prior knowledge of BLUF, entity echoing, schema markup, or fan-out queries — the items explain them in plain English. The checklist is most useful when you can hand individual items to the right person on your team (writer, developer, ops) without translation.