The AI Search Readiness Framework

Why readiness needs a framework

Over the past two years the interface of search has changed. Google answers a large share of queries with an AI Overview before the first result. ChatGPT, Perplexity and Gemini give buyers a shortlist of suppliers without a results page at all. Marketing teams know this is happening; most have no way of knowing whether their own site is ready for it.

“Ready” is not vague. Generative engines retrieve pages, judge them, and synthesise an answer. Each step depends on things you can audit: whether the engine can read you, whether you answer the question, whether the wider web vouches for you, and whether you're tracking the result. This framework turns those four stages into 24 yes/no checks you can score in an afternoon.

It's the same framework we use in client audits and the same structure as The Shortlist System. Score honestly, then fix the lowest-cost gaps first.

Stage 1 — Foundation (6 checks)

Foundation asks one question: can every engine crawl, render, index and understand your site? Score one point per “yes”.

1. Key pages render without JavaScript. Fetch a service page with JS disabled; the main content should still be present. AI crawlers are less tolerant of client-side rendering than Googlebot.
2. AI crawlers are not blocked unintentionally. Check robots.txt for GPTBot, PerplexityBot, ClaudeBot, Google-Extended and CCBot. Block deliberately or not at all.
3. Index coverage is clean. In Search Console, “Indexed” pages should match the pages you want indexed; “Crawled – currently not indexed” should be small and explainable.
4. Core Web Vitals pass on mobile for your main templates. Slow pages are retrieved less and trusted less.
5. Organization schema is present on every page, with name, URL, logo and sameAs links to your official profiles.
6. Business facts are consistent — name, description, location and category identical across your site, Google Business Profile, LinkedIn, directories and press. Inconsistency makes a model hedge or omit you.

Stage 2 — Answers (6 checks)

Answers asks: for each question a buyer might put to an assistant, do you have one clear page that answers it in the shape the engine expects?

1. You have a written list of buyer questions — at least 30, across discovery (“what is…”), comparison (“best… for…”) and decision (“how much does… cost”).
2. Each question maps to exactly one page. No question is unanswered, and no two pages compete to answer the same one.
3. Answer pages lead with the answer. A question-shaped heading followed within 60 words by a direct, complete answer.
4. Answers are specific. Numbers, named methods, concrete comparisons and dated statements, not “industry-leading solutions”.
5. Structure matches format. Lists for lists, tables for comparisons, numbered steps for processes.
6. Schema mirrors content. FAQ, HowTo, Article or Service schema that exactly matches what's visible on the page.

Most sites score 1–2 here. It's the stage with the highest return on effort — see our AEO Playbook for the page patterns.

Stage 3 — Trust (6 checks)

Trust asks: does the wider web vouch for you in the places engines check before naming a business?

1. You appear on the sources engines cite for your category. Ask Perplexity five category questions and note the cited domains. Are you listed, reviewed or mentioned on them?
2. Your review profile is healthy — recent, consistent, responded to — on Google and at least one category-relevant platform.
3. You're included in comparison and “best of” articles in your category, not just your own site.
4. Referring domains are relevant. Links from sites in or adjacent to your industry outweigh generic directories.
5. Your experts are visible. Named authors with bios, credentials and profiles; models weight content from identifiable people.
6. Press and third-party mentions describe you accurately — the same name, category and positioning as your own site.

Stage 4 — Measurement (6 checks)

Measurement asks: can you see whether any of this is working?

1. You test your buyer questions in AI assistants on a schedule — monthly at minimum — and record the answers.
2. You track mention rate: the percentage of tracked questions where you're named.
3. You track position and share of voice against named competitors.
4. You log cited sources so you know which pages (yours and others') engines rely on.
5. You monitor AI Overview inclusion for your Google queries, alongside impressions and clicks in Search Console.
6. AI visibility is in the same report as organic rankings and enquiries, so trade-offs are visible.

Our measurement guide gives a monthly process and a template.

Scoring and what to do with the result

Two patterns are common. Technically strong sites score high on Foundation and low on Answers and Measurement — the fix is content structure and a tracking habit, both inexpensive. Content-heavy sites score high on Answers and low on Trust — the fix is third-party presence, which takes longer. Either way, the framework tells you where the next point comes from.