Human oversight, documented — automatically

The EU AI Act asks organisations that deploy AI to keep humans in control and keep records of what the AI did — especially for higher-risk uses. For a small business, that usually means: know what your agents can do, be able to stop them, review risky decisions, and keep logs.

Halto gives you exactly that, as a side effect of keeping your agents safe:

• Written safeguards (your rules), versioned with who changed what and when
• Human approval for risky actions, with who decided, how and how fast
• An emergency stop, and a record of every time it was used
• An append-only, redacted log of every attempted action and the reason for each decision
• Retention you control (up to 12 months, depending on the plan) and one-click deletion

Not legal advice. Whether and how the AI Act applies depends on your use case and risk category.