Download the one-pager as a PDF here.

AI search is not just about ranking pages anymore. It is about making facts easy to retrieve, cite, and reuse. That usually means doing less clever keyword work and more plain-language work.

GEO stands for Generative Engine Optimization: the practice of making a page’s identity, answers, and evidence easy for an AI system to find and quote.

The useful shape is simple: one clear identity, one clear answer, and one piece of evidence per page. Models tend to be better at repeating facts when those facts are stable and easy to extract.

A few technical terms on the list below:

• JSON-LD is a machine-readable data block (usually in a script type="application/ld+json" tag) that tells tools what a page is about and who is behind it.
• @id is a unique web address you assign to a topic so machines can recognize it across pages instead of guessing.
• sameAs links that topic to your official profiles and sources elsewhere (for example your homepage, GitHub, or an organization record).
• /llms.txt is a short text file at the site root that gives AI tools a plain-language map of what your site covers and where the best pages are.

Here are the practical moves I would make first:

• Pick one canonical name and role for the brand, then reuse it everywhere.
• Use JSON-LD for Organization, Person, Service, Product, FAQ, and Article where relevant.
• Set @id, sameAs, author, datePublished, dateModified, and citation for the same reason.
• Publish a /llms.txt file with a short summary, key pages, and Markdown versions.
• Write sections that answer specific questions in one to three sentences.
• Link to primary evidence: datasets, benchmarks, changelogs, patents, or public work.
• Show authors, dates, methods, and what changed.
• Test across ChatGPT, Gemini, Claude, Perplexity, and Grok, then track what each model cites.

The one-pager covers the same approach with starter JSON-LD and /llms.txt snippets.