Kiteworks has announced the general availability of Agent and Human Error Prevention, extending its data control plane to outbound email to identify potentially harmful mistakes before sensitive information is sent to the wrong recipient.

The new capability is designed to assess multiple signals as a message is composed, including recipient type and volume, identity matching, attachments and domain validity.

The initial release includes three policies covering reply-all and BCC protection, send-to-self detection and domain typo detection. Administrators can configure policies as off, on or report-only, while user decisions and policy events are recorded in Kiteworks’ existing audit trail.

The capability operates within the customer’s environment, with email metadata not leaving the platform.

“The control plane for secure data exchange has always been about governing every send, share, receive, and use of sensitive data, for humans and agents alike,” said Tim Freestone, chief strategy officer at Kiteworks. “Misdirected email has sat outside that governance for years because it isn’t a data-pattern problem, it’s an intent problem. And intent is exactly what a policy engine that already knows the sender, the recipient and the context is positioned to catch.

"AHEP isn’t a new tool bolted onto the inbox. It’s a new module of the same Data Policy Engine we use to govern file sharing, MFT, APIs and agent access, extended to the channel with the least room for a scanning-after-the-fact approach."

Agent and Human Error Prevention is available for the Kiteworks Web App and Outlook Classic, including Outlook 2016 and later. Future phases are expected to introduce machine-learning-based behavioural analysis.

---