We didn't ask them to be case studies. We asked them to run a scan.
Every future case on this page will come from a real scan against a real staging or production AI agent. Company names, agent identifiers, and specific findings are anonymised where the customer hasn't consented to attribution — the risk scores and remediation timelines are not. Until then, this page is a scaffold.
What we actually found —
and what happened next.
Each case below will cover one scan, one finding worth talking about, and the remediation timeline that followed. No composite numbers, no rounded-up percentages. Today they are placeholders.
Additional scans are expected with organisations in insurance, enterprise SaaS, and public sector procurement under active NDA. Ask your design-partner contact for a live reference call rather than a written case study until this page graduates from scaffold.
From first scan to
verified fix — the real timeline.
This is the intended timeline for every case study that will appear here. Durations are medians from the product spec — not yet populated with partner data.
Findings pending design-partner beta — scaffold only — timeline durations above are spec medians, not yet partner-verified.
Your agent hasn't been scanned this way yet.
Every future case study on this page will start the same way every early access account starts — a free first scan, no sales call required. Join the design-partner waitlist to be notified when the first findings publish.
Customer names withheld by request when published · Findings and figures will be real, unaltered scan data · Findings pending design-partner beta — scaffold only