Service 08 / 14
Clear enough for an assistant to quote.
ChatGPT, Gemini, Perplexity and Google’s AI answers read the web and quote what they can understand. We make your business easy to understand and easy to quote.
In short
What this covers
An assistant answering “who repairs heat pumps in Laval?” needs three things from your site: to know what the business is, to know what it does and where, and to find a sentence it can trust enough to repeat. Most sites bury all three.
AEO is mostly good structure, applied strictly: one plain definition, services and areas named the same way everywhere, questions answered in the first lines, the business described for machines, proof that can be checked. We score it, fix it, and keep the score honest.
- Usually for
- Businesses asked about by name in assistants, Companies whose services are hard to describe in one line, Sites with good content that no assistant seems to use
The work
What we do
-
Entity
Who the business is, stated once and the same way everywhere: name, type, founder, profiles, contact. Marked up as an Organization or LocalBusiness.
-
Services and areas
Each service and each area named consistently in text, navigation and structured data.
-
Direct answers
Openings that answer, questions as headings, short paragraphs that stand alone when quoted.
-
Machine-readable context
JSON-LD graph, an llms.txt summary, clean sitemaps.
-
Proof
Facts an assistant can check: dates, statuses, sources. Nothing invented, so nothing to retract.
-
Readiness score
Eight areas scored from what is on the page, each point with its evidence.
Architecture
How the system fits together
The AI Search Readiness score is computed from your pages, not estimated. Observed answers from assistants are a separate, dated sample.
-
Read
Your pages and structured data, as the audit reads them.
site/server/growth/audit -
Score
Entity, services, locations, answers, structured data, trust, proof, completeness: 0–100 each, with the evidence.
site/server/growth/aeo.js -
Fix
Definitions, answers, markup and proof added where the score says they are missing.
Content model -
Observe
A fixed list of your customers’ questions asked to assistants on a schedule, answers stored with their date. An observation, never a guarantee.
Planned
Status
What runs today, and what waits for access
Live means it runs now. Integration-ready means the code, the data model and the tests exist and it starts once your account or a platform approval is connected. Planned means not running yet: designed, or built and tested but not yet connected to a live endpoint.
-
Live
AI Search Readiness score
Computed on any site from the audit at /audit/ (the “AI search readiness” area), and in full by the AEO engine.
-
Live
Entity graph and llms.txt
This site publishes both, generated from its content.
-
Planned
Observed answers from assistants
Sampled questions, dated, labelled as observation, where each assistant’s terms allow it.
Measurement
What we measure, and where the number comes from
- Readiness score per area
- Computed from the pages, with the evidence for every point.
- Visits and leads from assistants
- The attribution engine treats ChatGPT, Perplexity, Gemini, Copilot and Claude as their own channel, from referrers and tags.
Handover
What you end up with
- A readiness report with evidence
- The definition, answers and markup fixed on the site
- An llms.txt summary
- A question list to observe over time
Limits
Where this isn’t the right fit
No one can guarantee that an assistant will cite you. We make it possible and easy, and we say so.
Assistants change their sources and behaviour often. We measure what we can see and date every observation.
FAQ
Questions about this
What is AEO?
Answer-engine optimization: making a business easy for AI assistants (ChatGPT, Gemini, Perplexity, Google’s AI answers) to understand and quote. It is mostly strict structure: a plain definition, services and areas named consistently, questions answered directly, and structured data. The audit scores it on any site.
Do you guarantee rankings, leads or AI citations?
No. Search engines, ad platforms and AI assistants decide what they show, and anyone who guarantees a position is guessing or buying something they shouldn’t. We build what we control properly, measure what happens with stated methods, and report it, including when it doesn’t work.
See how readable your site is for an assistant.
The audit scores AI search readiness in about ten seconds.


