/ Products · Geneio
Search is becoming answering. Most sites are not ready for it.
When someone asks an assistant for a recommendation, the answer is composed rather than listed — and a site tuned for search rankings can be entirely absent from it. Geneio scores how ready a domain is to be read, understood and cited by an answer engine, shows precisely where it falls short, and tracks the score as you fix it.
/ Why it matters
A ranking is not a citation.
Two decades of search optimisation taught organisations to compete for position on a page of links. Answer engines do not produce a page of links. They produce a paragraph, assembled from sources they judged legible and trustworthy enough to draw on — and being ranked first for a keyword does not put you in it.
Different unit of success
The question is no longer where you rank. It is whether you were used as a source, whether you were described accurately, and whether you were named at all.
Different readership
The first reader of your site is now a machine deciding whether your content can be parsed, attributed and safely quoted. Most sites were never written for that reader.
No feedback loop
Rankings are measurable and familiar. Whether an assistant considers your site a credible source is not something most organisations have any instrument for.
/ How it works
A domain, a few keywords, a score you can act on.
/01
Content
Whether the material answers questions directly, is structured for extraction, and states things a machine can lift without distorting them. The heaviest weighting, because it is the thing most often wrong.
/02
Technical
The mechanics that decide whether a crawler can reach and parse you at all — transport security, markup validity, crawl directives, and whether AI crawlers are admitted deliberately rather than by accident.
/03
AI readiness
Structured data, resolvable entities, and machine-readable context — the difference between a page a model can quote confidently and one it will paraphrase badly or skip.
/04
Authority
The corroboration that decides whether a source is trusted enough to be drawn on when an assistant composes an answer.
/ GEO and SEO
The same goal, a different instrument.
Generative engine optimisation does not replace search optimisation; it addresses a channel that search optimisation was never designed for. The distinction is worth stating plainly, because the two are routinely confused.
Measured by
SEO: rankings, sessions, clicks·GEO: mentions, citations, accuracy of description
Optimised for
SEO: keyword match and link equity·GEO: context, structure and demonstrable authority
Surface
SEO: a page of ranked links·GEO: a composed answer that may cite nobody at all
Read by
SEO: a crawler indexing for retrieval·GEO: a model deciding whether you are quotable
/ Evidence
We run it on ourselves. You can check.
This site is built to the standard Geneio measures, and the evidence is public rather than asserted. Every page carries a structured entity graph. The questions on our approach page are marked up so an answer engine can quote them correctly. Our llms.txt states what we do and — more usefully — what we will not claim, so a model summarising us has the caveats in hand. Our robots file admits AI crawlers deliberately.
That is the argument in full. A firm selling generative engine optimisation ought to be legible and citable itself. Open the files and see.
/ Built on
A real application, not a landing page.
Geneio is a deployed product with accounts, sessions and persistence — the same engineering standard we apply to client work.
Frontend
React 19·Tailwind CSS·React Router·Responsive with light and dark themes
Backend
FastAPI·Pydantic validation·Async data access
Data
MongoDB·Score history retained per domain
Accounts
JWT sessions with refresh·Bcrypt password hashing·Email verification and reset·Lockout on repeated failures
Operations
Containerised·Kubernetes with liveness and readiness probes·Degrades rather than crashes when a dependency is unavailable
/ Where it goes
Named plainly, so a pilot starts honestly.
These are in development, not in the product. We would rather you heard that from us than discovered it in week two.