That's ~€15.80 per client per month at 5 clients — less than agencies pay for white-label
SEO report widgets, for the conversation that actually gets you fired if you fumble it.
Questions agencies ask
"My platform says it migrates models automatically."
If everything you run lives on Make, GoHighLevel, or Botpress: correct, you’re mostly covered, and we’d rather tell you that than sell you a report. This desk is for raw-API, Vapi, Retell, and n8n stacks, where nobody migrates anything for you.
"Will the report say my client’s agent still works?"
No, and be suspicious of anyone whose report would. We publish what standardized replay can honestly measure: cost deltas in euros, parameter failures, schema flags, probe-outcome consistency, disclosure observed or not observed. That’s the evidence for your judgment call.
"Is this an AI-Act compliance certificate?"
No. The disclosure probe documents observed behavior with a sealed transcript. Useful when procurement asks; not a legal assessment.
"What do you need from us?"
A config bundle per agent: model and parameters, system prompt, tool schemas, optionally the live widget URL — with your client’s consent, attested at registration. No end-user data, no eval corpus, no sandbox integration.
"Who sees our clients’ configs?"
Standard DPA at signup; configs are processed solely to produce your reports; model-provider APIs are named sub-processors; deletion on request.
"Why should the report carry your name?"
It doesn’t have to — Variant A is fully agency-branded, our name appears nowhere client-facing. Some agencies find the co-signed "Independently verified by Stratt Labs" variant closes the client conversation faster. Pick per client, switch anytime.