Get found by AI, with a site assistants can read.
More and more customers ask an AI assistant instead of searching Google. We prepare your site so assistants can read it, understand it and cite it, and we measure the result.
An assistant cites only what it can read.
AI assistants answer by combining the pages they can download and understand. A site that is only built in the browser looks empty to them: the product site of a company we worked with had zero pages indexed. A heavy site, with no rules for AI crawlers and no structured data, simply isn't in the answers. And hardly anyone knows what assistants say about their business today.
Measure first, then act.
- 01
What they answer today
We ask the main assistants the questions your customers would ask, always the same ones in the same way, and archive the answers. That's the baseline: who gets cited, in what words, from which sources.
- 02
Can the site really be read?
We check that every page arrives complete from the server, without needing JavaScript, and measure agent readiness with an independent tool, isitagentready.com.
- 03
Signals for agents
Explicit rules for AI crawlers in robots.txt, Content Signals to decide what can be indexed, used in answers or used for training, an llms.txt index, a Markdown version of every page and JSON-LD structured data.
- 04
Pages that answer
Assistants cite pages that answer a precise question. We work with you on services, cases and FAQs written to answer, with real numbers and sources.
- 05
Measure again
We repeat the scan and the same questions to the assistants and compare with the baseline. Then we repeat them over time, because assistants change.
A ready site, and the numbers to prove it.
On WordPress sites the agent layer lives in a plugin of ours, CF - Agent Discovery: theme and content stay as they are, and it can be removed without a trace.
Those who live on enquiries.
Service firms and companies
Anyone who gets customers from a question like "who does this in Rome?" and wants to be in the answer.
WordPress sites
The CF - Agent Discovery plugin adds llms.txt, Markdown versions and Content Signals without touching the theme.
Sites built in the browser
React or similar applications that search engines see as empty: we bring them back to complete pages generated on the server.
Content-heavy portals
Archives of articles and listings that need ordering and making readable before they can be cited.
We do it on ourselves first.
This site has its own llms.txt, a Markdown version of every page and structured data on services, cases and articles. We apply the same method, survey, intervention and new survey, to clients' sites. And because we also manage servers and Cloudflare, we can act on everything that decides whether a page reaches a crawler or not.
Where we've done it before.
A funding portal ready for AI assistants.
Clean-up first, then signals for agents: rules for AI crawlers, Content Signals, an llms.txt index. A dedicated plugin, measured before and after with an independent tool.
Two websites, one application, and finally visible to search engines.
One Laravel and Filament app serves two domains from the same code and database: complete server-side HTML, JSON-LD with the existing identifiers, press review and forms migrated from WordPress, a server the client can take over.
Measurement you can trust.
Tracking, GA4, consent management and technical SEO put in order: no conversions counted twice, no tags before consent, no pages competing for the same searches.
Before we start.
Can you guarantee ChatGPT will cite us?
No, and nobody can honestly do so: answers depend on the assistant, the question and what others write. We can guarantee your site is readable, and measure how answers change before and after.
Do we need to rebuild the site?
Usually not. On WordPress our plugin plus some work on the content is enough. The site only needs rebuilding if its pages are built in the browser today and crawlers see them as empty.
Should we block AI bots?
It depends on what you want. With Content Signals you decide separately whether content can be indexed, used in answers or used to train models. We discuss it with you before writing the rules.
What about advertising inside assistants?
Where assistants show paid placements, we assess them alongside the campaigns you already run. But the foundation is still a site assistants can read: that's what makes the ad credible too.