Technical / GEO · 2026
Is Your Website Ready for SEO and AI? A Checklist
A website is ready for SEO and AI (GEO) when three layers are in place together: technical health (mobile-friendliness and Core Web Vitals — LCP ≤ 2.5 s, INP ≤ 200 ms, CLS ≤ 0.1), structured content (a clear answer in the first 300 words, question-style headings, tables, FAQ and schema), and open access for AI bots (robots.txt not blocking crawlers such as GPTBot, ClaudeBot, PerplexityBot and OAI-SearchBot); the 12-point checklist below lets you see in about 10 minutes which layer your site is missing.
A website being "ready" no longer means only ranking on Google. In 2026 a significant share of users get their answer straight from ChatGPT, Perplexity and Google's AI Overviews box. So readiness has two directions: classic SEO (ranking in search engines) and GEO — Generative Engine Optimization (being the cited source inside AI answers). This page is a practical 12-point checklist to measure both at once. If you are planning a new site, our corporate website step-by-step guide helps you build this checklist in from the start.
Are SEO and GEO the same thing? What is the difference?
They are related but not identical. SEO aims to appear higher on a search results page (SERP); the reward is a click. GEO aims to be named as a source inside the answer an AI generates; the reward is visibility and trust. The good news: the fundamentals largely overlap. Fast, mobile-friendly, structured and honest content both ranks on Google and gets cited by AI. The checklist below gathers this common ground into three layers.
The 12-point checklist (about 10 minutes)
Tick each item off on your own site. The "how to test" column shows, for most of them, a free tool or a simple check you can do from your browser.
| # | Layer | Check | How to test | Target |
|---|---|---|---|---|
| 1 | Technical | Mobile-friendliness | Search Console → Mobile Usability; open on a phone | No errors, readable size |
| 2 | Technical | Page speed / Core Web Vitals | PageSpeed Insights — the "real user" section | All three metrics green |
| 3 | Technical | HTTPS + valid SSL | Padlock icon in the address bar | Padlock present, no warning |
| 4 | Technical | Indexability | Search Console → Pages; robots.txt must not block the crawler | Key pages indexed |
| 5 | Content / GEO | Clear answer in first 300 words | Is there a one-sentence answer at the top of the page? | Yes |
| 6 | Content / GEO | Question-style headings (H2/H3) | Are headings phrased as "what is / how / how much"? | Most headings questions |
| 7 | Content / GEO | Data as tables / lists | Is there at least one structured table/list on the page? | Yes |
| 8 | Content / GEO | FAQ section | Are there short answers to real user questions? | 3+ questions |
| 9 | Content / GEO | Structured data (schema) | Scan with Google's Rich Results Test | Article + FAQPage valid |
| 10 | AI access | AI bots allowed in robots.txt | Open site.com/robots.txt; are GPTBot / OAI-SearchBot / ClaudeBot blocked? | No block |
| 11 | Trust / E-E-A-T | Author + company identity | Are a named author, contact and "about" visible? | Visible + schema |
| 12 | Trust | Freshness ("last updated") | Does the content carry a date and current data? | Dated + current |
Core Web Vitals: which values count as "good"?
Core Web Vitals are Google's three metrics for user experience and are assessed for 75% (p75) of real users. In other words, three quarters of visits must have a good experience for your page to "pass".
| Metric | What it measures | "Good" threshold |
|---|---|---|
| LCP | Loading — time for the largest content to appear | ≤ 2.5 seconds |
| INP | Responsiveness — replaced FID in March 2024; measures every interaction | ≤ 200 ms |
| CLS | Visual stability — unexpected layout shifts | ≤ 0.1 |
How do you check AI bot access?
For AI engines to cite your site, they must first be able to reach your pages. This access is managed by the yourdomain.com/robots.txt file. The critical distinction: some bots crawl for model training, others for live search/citation. Blocking the search bot drops you not from training but directly from today's AI answers.
- GPTBot (OpenAI) — crawls for model training. Blocking it removes you from future training; it does not directly affect today's answers.
- OAI-SearchBot (OpenAI) — ChatGPT's search/citation bot. Block it and you drop from ChatGPT search citations — a direct visibility loss.
- ClaudeBot / Claude-SearchBot (Anthropic) — training and search/citation crawlers respectively.
- PerplexityBot (Perplexity) — crawls to show sources in Perplexity answers.
- Google-Extended (Google) — controls Gemini/AI training permission; it is separate from the classic Googlebot and does not affect normal search.
Rule of thumb: if you want AI visibility, do not close any of these bots with Disallow; also make sure any CDN/firewall (e.g. Cloudflare) "block AI bots" option is off — a block there will stop the bots even if robots.txt is open. (Source: honeyb.ai, 28 May 2026.)
What does it take to get cited by AI engines?
Once access is open, the next step is content in the shape AI likes to cite. The academic GEO study led by Princeton (KDD 2024) measured that adding citations, statistics and credible sources can raise a source's visibility in generative-engine answers by up to 40%. In practice, what helps most:
- Give a clear answer in the first 300 words — AI judges relevance largely from the opening content.
- Write headings as questions; get close to the exact phrasing users ask the AI.
- Add concrete data and statistics + a source; a dated, anchored figure gets cited over a generic claim.
- Add an FAQ section — disproportionately powerful for conversational queries.
- Make author and company identity visible (E-E-A-T); a named expert outperforms anonymous content.
- Keep it current; a "last updated" date and refreshed data stand out for time-sensitive queries.
Note: Google's AI Overviews box appears on roughly 30–40% of queries by some measures (source: Enrich Labs GEO guide, 2026), showing that GEO is now a mainstream visibility channel rather than a niche. These items are also good user experience and classic SEO — so one effort feeds both channels.
If your score is low, where should you start?
Order matters. First, access and the technical base (items 1–4 and 10): without these, even the best content is invisible. Then content structure (items 5–9): without deleting your existing pages, you can improve most of them quickly by turning headings into questions and adding a clear answer to the first paragraph. Last, the trust layer (items 11–12) is an investment that builds over time. As you budget for this work, reviewing how we calculate value for money in web design and the current web design price ranges will make it easier.
Frequently asked questions
What is the difference between SEO and GEO?
SEO aims to appear higher on a search results page (Google); GEO aims to be cited as a source inside the answers of generative engines such as ChatGPT, Perplexity and Google AI Overviews. The fundamentals (speed, mobile-friendliness, structured and honest content) largely overlap, so it is possible to optimize for both at once.
Which Core Web Vitals values are good enough?
As of 2026 the "good" thresholds are LCP ≤ 2.5 seconds, INP ≤ 200 ms and CLS ≤ 0.1 for 75% of real users. INP is the responsiveness metric that replaced FID in March 2024. You can see the values in the "real user" section of Google PageSpeed Insights.
How do I allow AI bots to crawl my site?
Do not block bots such as GPTBot, OAI-SearchBot, ClaudeBot and PerplexityBot with Disallow in robots.txt; also, if you use a layer like Cloudflare, make sure its "block AI bots" option is off. Blocking OAI-SearchBot in particular drops you from ChatGPT's search citations.
How do I improve the chance of being cited by AI?
Give the clear answer in the first 300 words, phrase headings as questions, add tables/lists and an FAQ, anchor your claims with dated statistics and sources, make author/company identity visible, and keep the content current. According to the academic GEO study, adding citations, statistics and sources can raise visibility by up to ~40%.
Do I need to add an llms.txt?
Not required for now, and its benefit is not yet clearly proven. Your priority should be opening AI bot access in robots.txt, technical speed and structured content; llms.txt is a lower-priority experiment after those.
Sources
- Google web.dev — "Core Web Vitals" (LCP/INP/CLS thresholds; p75 measurement) — accessed/published: 2026
- corewebvitals.io — "What Are the Core Web Vitals? (2026)" (INP replaced FID in March 2024) — accessed/published: 2026
- Honeyb — "AI Crawler User Agents Reference (2026)" (GPTBot/OAI-SearchBot/ClaudeBot distinction) — accessed/published: 28 May 2026
- Aggarwal et al. — "GEO: Generative Engine Optimization" (Princeton/Georgia Tech), ACM SIGKDD 2024 — accessed/published: KDD 2024
- Enrich Labs — "Generative Engine Optimization (GEO): The Complete 2026 Guide" (AI Overviews prevalence; first-200-words principle) — accessed/published: 2026