Technical foundation is the layer everything else in GEO sits on. If AI crawlers can't reach your pages, parse the HTML or understand what each page is about, none of your content or off-site work will translate into citations. The good news: the technical bar for GEO is lower than enterprise SEO makes it sound. A small B2B SaaS team can cover the essentials in a single sprint.
What is technical GEO?
Technical GEO is the set of site-level decisions that make your pages accessible and intelligible to AI crawlers - things like robots rules, server-side rendering, structured data, internal linking and speed. It overlaps heavily with traditional technical SEO, but the priorities differ. AI tools care less about Core Web Vitals scores and more about whether the answer to a question can be extracted from your HTML without running JavaScript.
How much does technical foundation matter?
More than people think, and less than agencies want to sell you. The first 20% of the work - making sure AI crawlers are allowed in, serving real HTML, shipping clean structured data - captures most of the value. Beyond that, technical effort has steeply diminishing returns compared to investing in citable content and off-site authority.
If you're a fast-moving SaaS with a modern stack, you've probably already done 70% of this without calling it GEO.
Each layer depends on the one below it
Which AI crawlers should I let in?
At minimum, allow the major AI crawlers in your robots.txt:
GPTBot- OpenAI / ChatGPT searchOAI-SearchBot- OpenAI's dedicated search indexClaudeBotandanthropic-ai- Anthropic / ClaudePerplexityBot- PerplexityGoogle-Extended- Gemini training and AI Overviews opt-inApplebot-Extended- Apple Intelligence
Blocking these crawlers is the single most common reason a SaaS site doesn't appear in AI answers. Check your robots.txt today. If you inherited a wildcard Disallow rule, you may be invisible to half the AI ecosystem without realising it.
Does my site need to be server-rendered?
For the pages you want AI to cite: yes, or close to it. Most AI crawlers do not execute JavaScript the way a modern browser does. If your pricing page, product page and key guides only render their content after a client-side JS bundle runs, AI tools will see an empty shell.
Use SSR, static generation or pre-rendering for any page that has citation potential. Single-page app shells are fine for authenticated product UI, but your public marketing and content surface should serve real HTML.
What structured data should I ship?
Four blocks of JSON-LD cover 80% of the GEO value:
- Organization on your homepage, with a thorough
sameAsarray pointing to LinkedIn, G2, Capterra, GitHub and your other verified profiles. - SoftwareApplication (or Product) on your main product page.
- FAQPage on any page with question-and-answer content.
- Article on every blog post and guide, with a real named author.
For the full breakdown of what to use and what to skip, see Schema markup that actually helps AI cite you.
Do speed and Core Web Vitals matter for GEO?
They matter, but as a hygiene factor. AI crawlers have limited time per site. If your pages take five seconds to respond, fewer get crawled and your fresh content takes longer to enter the AI index. Aim for a server response time under 600ms and you've done enough. Obsessing over a 95+ Lighthouse score is not where the marginal GEO win lives.
What about sitemaps and llms.txt?
A clean, accurate sitemap.xml is still the best way to tell crawlers what exists on your site. Auto-generate it from your routes so it stays in sync. Link it from your robots.txt with a Sitemap: directive.
llms.txt is a proposed convention for pointing AI tools at your most important pages. It is not yet honoured by the major AI tools, but it costs nothing to ship and signals intent. Treat it as a nice-to-have, not a foundation.
A one-week technical foundation checklist
- Day 1: Audit
robots.txt. Explicitly allow GPTBot, ClaudeBot, PerplexityBot, Google-Extended. - Day 2: Confirm your top 20 pages render real HTML server-side. View source, not inspect.
- Day 3: Ship Organization and SoftwareApplication JSON-LD.
- Day 4: Add FAQPage and Article schema to your highest-traffic pages.
- Day 5: Generate a dynamic
sitemap.xmland reference it fromrobots.txt. - Day 6: Check server response times. Anything over 1s gets a ticket.
- Day 7: Validate every schema block in Google's Rich Results Test.
That is the technical foundation. Once it's in place, your time is far better spent on the next three pillars: on-site optimisation, off-site authority and monitoring.