The data service
Measurement that survives scrutiny.
The measurement layer installed during implementation and operated as part of monitoring. It records which identified AI services requested which pages, when, and whether access was allowed.
How it is installed
Three reporting paths.
Which one you use depends on what you control. They are not exclusive, and most brands end up with the first two.
At your CDNAvailableA worker at the edge identifies AI retrieval requests and reports them.
Deployed to Cloudflare or Fastly in front of your existing origin. It sees every request, including the ones that never reach your application, and it does not change what your users are served. This is the path most brands take first because it needs no change to the site itself.
At your web serverAvailableMiddleware or log shipping from the origin you already run.
Where a CDN worker is not an option — an origin behind someone else’s infrastructure, or a hosting arrangement you do not control — the same events can be derived from server middleware or from access logs you already retain.
From the AI platformDepends on the platformThe platform reports what it grounded on and cited.
The richest signal, and the only one that can distinguish content that was merely fetched from content that actually entered a model’s context. It requires the AI platform to report, which is a matter for the standard and its adoption rather than something a brand can deploy unilaterally.
What it shows
And what it does not.
What an edge record establishes
- An identified service requested a specific URL.
- When the request happened.
- Whether your policy allowed or blocked it.
- What your configuration actually did, as opposed to what it was meant to do.
What it does not establish
- That the content was used for training.
- That it entered the context of any particular answer.
- That it was cited, displayed, or seen by a person.
- That it caused a visit, a conversion or a sale.
Prompt monitoring is a separate observation of what AI systems say and cite. We report both, and we do not join them into a causal claim.We record what crossed, not what influenced.
Status
What you can have today.
Conformant Content Telemetry ingestion atapi.openattribution.org, which we operate under contract from OpenAttribution. Live, with an end-to-end canary running against it every fifteen minutes.
Edge telemetry integration at your CDN or origin, installed as part of an implementation engagement, with conformant export.
A self-serve dashboard with domain verification, a free bounded reporting baseline for site owners, prompt monitoring and a discrepancy workflow. Today these are delivered through an engagement rather than by signing up.
Your data
Exportable, in a format we do not own.
Events are recorded againstContent Telemetry, an open standard governed by SPUR. You can export your data and take it to any other conformant service.
Our value is the implementation, the evidence and the workflow. A measurement service that retains customers because leaving is painful is not worth building, and would not be credible as measurement in any case.
Start with the audit.
A fixed-scope review of one brand, product category and set of priority markets. You get an evidence-based baseline and a prioritised plan with owners and costs — not a report that sits on a shelf.