SEO and GEO: Build Websites That Search and AI Can Cite

In short: A website now has to be found by search engines and understood and cited by AI answers. Google says the same SEO fundamentals apply to its AI features, so GEO (Generative Engine Optimization) starts with crawlable pages, fast performance and clear structure in the codebase. On top of that, you make your facts easy to quote: consistent entity details, structured data that matches the page and answer-shaped content.

In early 2025 I wrote about why SEO belongs in website development, not after it. The core point still stands. What changed is where people get their answers.

A growing share of discovery happens inside an AI answer. Someone asks ChatGPT, Gemini, Perplexity, Grok or Google Search a question, gets a summary and maybe clicks one or two cited sources. Google reports that AI Overviews has over 2.5 billion monthly active users and AI Mode has passed one billion (Google, 2026).

So the question I ask on every website project is no longer only "will this rank?" It is "will this be found, understood and cited?" That is SEO plus GEO, and both start in the codebase.

SEO is still the foundation

Here is the part many people get wrong about GEO. Google's own guidance says the same SEO best practices apply to AI Overviews and AI Mode, and that there are no additional technical requirements to appear in them (Google Search Central). A page has to be indexed and eligible for a snippet first.

That means the unglamorous engineering work matters more, not less:

• Crawlable pages. Important content is in the HTML as text, not hidden behind client-side rendering, logins or images.
• Clean structure. One H1, logical headings, descriptive URLs and internal links that reflect how your services actually relate. We wrote more about this in why site architecture is the hidden engine of growth.
• Correct sitemaps and robots rules, including at the CDN and hosting layer, where crawlers are often blocked by accident.
• Canonicals, hreflang and redirects planned before a redesign goes live, not after traffic drops.

None of this is new. It is just easy to skip when a team treats SEO as a plugin.

Performance is a feature you build, not a setting

Google's Core Web Vitals give concrete targets: Largest Contentful Paint within 2.5 seconds, Interaction to Next Paint of 200 milliseconds or less and Cumulative Layout Shift of 0.1 or less, measured at the 75th percentile of real page loads (web.dev).

The shift worth noting is INP. It replaced First Input Delay in 2024 and it penalises heavy JavaScript that blocks the main thread after the page looks loaded. A caching plugin does not fix that.

What works in practice:

• Set a performance budget per page type and fail the build when it is exceeded.
• Serve modern image formats at the right size, and reserve space for images and embeds so the layout does not jump.
• Ship less JavaScript. Render on the server where you can and hydrate only what is interactive.
• Measure in the field, not only in Lighthouse. Lab scores are a guide; real users on mid-range phones are the test.

Fast pages help people first. They also help every crawler, including the ones that feed AI answers, fetch and parse your content reliably. More on this in how website speed affects SEO and UX.

GEO: make your facts easy to quote

AI systems assemble answers from many sources. Your job is to be the clearest, most consistent source for the facts that concern you. This is where GEO adds a layer on top of classic SEO.

Clear entity facts. Say plainly who you are, what you do, where you are based and how to contact you. Use the same company name, address and service names everywhere: your site, your Business Profile, LinkedIn and directories. Contradictions make you harder to cite.

Structured data that matches the page. Organization, LocalBusiness, Product, FAQPage, Article and BreadcrumbList schema help machines read context. Google is explicit that structured data must match the visible text, and that no special schema is required for AI features. Treat schema as a precise label, not a trick.

Answer-shaped content. Write sections that each answer one real question, with the answer near the top. Use specific nouns, dates and numbers you can stand behind. Vague marketing copy rarely gets quoted.

Original information. Pricing logic, process descriptions, technical documentation and comparisons you have actually done. If your page says the same thing as a hundred others, there is no reason to cite yours.

About llms.txt

llms.txt is a proposed convention, first published in 2024: a Markdown file at /llms.txt that gives AI agents a short summary of a site and links to clean, LLM-friendly versions of key pages (llmstxt.org).

I would describe it carefully. It is an emerging convention, not a standard, and not every AI system reads it. Google says you do not need new machine-readable or "AI text" files to appear in its AI features. My view: on a well-built site it costs little, it does no harm and it is useful for agents reading documentation. Add it after the fundamentals, not instead of them.

A practical checklist for 2026

When we build or rebuild a site, this is the order I care about:

1. Indexability: rendered HTML, sitemaps, robots and CDN rules checked.
2. Core Web Vitals within Google's "good" thresholds on real mobile data.
3. A content structure built around the questions your customers ask.
4. Schema for your organisation, services, articles and FAQs, validated and matching the visible text.
5. Consistent entity facts across your site and external profiles.
6. llms.txt and Markdown versions of key pages, where they make sense.
7. Search Console and analytics wired in from day one, so AI-driven traffic shows up in reporting.

Then keep it maintained. Content drifts, plugins add weight and redirects break. SEO and GEO are part of the software, so they need the same care as the software. That is why ongoing support and maintenance belongs in the plan from the start.

How LabForty approaches it

Technical SEO is part of every website we deliver. Our website development scope covers page structure, metadata, indexing and performance, and for redesigns we plan which content, URLs and integrations need to be kept or changed. Rankings and traffic still depend on competition, content and ongoing work, and we say so up front.

You can see how we structure content-heavy sites in our case studies, for example the fund catalogue and product pages on the Themes ETFs website.

FAQ

What is the difference between SEO and GEO?

SEO makes pages crawlable, indexable and relevant so they rank in search. GEO builds on that so AI systems can understand your content and cite it in generated answers. GEO does not replace SEO; it depends on it.

Do I need special files or schema to appear in Google's AI Overviews?

No. Google says the same SEO best practices apply and there are no additional technical requirements. Pages must be indexed and eligible for a snippet (Google Search Central).

Should my website have an llms.txt file?

It is an optional, emerging convention. It can help AI agents read documentation, but it is not a standard and not every system reads it. Add it after the fundamentals are in place.

Which performance targets should a new website meet?

Google's "good" Core Web Vitals thresholds: LCP within 2.5 seconds, INP of 200 milliseconds or less and CLS of 0.1 or less, at the 75th percentile of real visits (web.dev).

Plan a site that search and AI can cite

If you want a website that people and AI systems can both find and trust, tell us what you are building. A rough idea is enough to start, and we reply within 24 to 48 hours on working days.