Security & Trust
Handlers act inside your systems. So the real question isn't just is the data encrypted — it's what can the AI actually do, and what stops it doing the wrong thing? Here's how that's controlled by design.
What the AI can — and can't — do
The baseline, done right
Take the next step
Whether you need the paperwork or a conversation about your architecture requirements, there's a path for where you are.
Get the documents
SOC 2 report and policy docs at trust.handle.work. Security questionnaires and penetration-test reports on request.
Request the documentsTalk to us about your architecture
A security review with an architecture walkthrough, audit-log examples, and named integration patterns — for the agentic-control detail we keep off the public page.
Book a security review