A first evaluation typically takes about three weeks and classifies observed behaviour across 8 measurement dimensions. Because your AI system changes with every release, the first run is a baseline rather than an endpoint — the comparison between runs is where the evidence earns its keep.

For the cloud analysis, no internal integrations are needed — we work with public content, document exports, and API endpoints. The on-premise deployment of Proof integrates via API with your existing AI infrastructure.

We operate under SOC 2 controls, encrypt data in transit and at rest, and never surface PII unprotected. On-prem deployments keep all content within your firewall.

We provide before-and-after metrics on risk reduction, time saved vs. manual review, and positive sentiment uplift—so you can calculate cost avoidance (fines, headcount, brand damage).

Yes. We support multi-language analysis and global channels, and you can configure per-region thresholds and per-brand rule sets.