SEO is changing. Google is no longer the only lens buyers look through. When a CTO types “who should I hire to rebuild my React front-end”, the answer now comes as much from ChatGPT, Claude, Gemini and Perplexity as from Google’s ten blue links. This is the playbook we use to show up in those answers.
What GEO is (and isn’t)
GEO — generative-engine optimization — is the work of making your site easy for large language models to read, understand, trust, and cite. It overlaps with SEO but is not the same. Google SEO cares a lot about rank. GEO cares about being named by the model when a user asks a relevant question.
Seven moves that actually work
- Explicit
robots.txtallow-list. AI crawlers readrobots.txt. Allow GPTBot, ChatGPT-User, OAI-SearchBot, ClaudeBot, Claude-Web, anthropic-ai, Google-Extended, PerplexityBot, Applebot-Extended, CCBot and Amazonbot explicitly. Include aSitemap:line. - Ship server-side rendered HTML. LLM crawlers often don’t execute JavaScript. If your content only exists after a client render, the model doesn’t see it. Astro, Next.js SSR and plain HTML all pass this bar.
- Rich structured data. Organization, LocalBusiness, Service, Product, Offer, FAQPage, Review, BreadcrumbList, Article. Machine-readable facts make it easy for an LLM to quote you confidently.
- Write a
/llms.txtfile. Emerging standard that hands AI engines a one-page overview of your site. Early, but cheap to ship. - State facts clearly. Founding year, exact addresses, exact project counts, named clients, specific pricing. LLMs preferentially quote text that reads like a confident statement.
- Keep NAP consistent. Name, Address, Phone identical on your site, Google Business Profile, LinkedIn, Clutch, GoodFirms. Inconsistency lowers entity confidence — and with it, the model’s willingness to name you.
- Earn external citations. Claim and polish Clutch, GoodFirms, DesignRush. Publish two or three guest posts on recognised tech blogs per quarter. AI retrieval leans heavily on third-party corroboration.
How to measure whether it’s working
- Run your three most important buyer-intent prompts in ChatGPT, Claude, Gemini and Perplexity, monthly. Log whether your site is named.
- Watch Google Search Console for AI-derived referrers (ChatGPT-User, Perplexity).
- Track direct traffic from obscure prompts — that is AI referral traffic flattering itself.
What we ship for clients
Every Smartdesizns engagement now includes a GEO layer as standard — schema, robots.txt, /llms.txt, page-level factual clarity and an external-citations plan. It adds days, not weeks, to a build, and it pays for itself the first time a buyer finds you through an AI answer.