Post by Island

42,755 followers

If you can't see inside an AI session, you aren't governing AI. You're guessing. Island enables AI by governing where work actually happens. Today Island integrates with the Claude Compliance API, from Claude, powered by Anthropic. Watch what that looks like. A user drops a draft financial workbook into Claude and asks for board-ready talking points. Island captures the whole session: - The file - The prompt - The tools it touched - The sensitive data inside it Island ties those to one identity alongside everything we already see from the workspace, network, and endpoint. Chats, uploaded files, and tool calls that never cross a network boundary are now in view, classified for PII, PHI, PCI, source code, and secrets. The more surfaces Island sees, the less any single AI interaction can slip past policy. It's available now. More from Avishai Winiwarter's blog: https://lnkd.in/e-n74gdG Michael Fey Dan Amiga Bradon Rogers Alon B. Daniel Kandel Roi Leibovich Wojtek Erbetowski Moshe Alvoer Adam Dąbrowski Kasia Kucharczyk Błażej Marcinkiewicz Michal Piotrkowski Uri Harduf Yahel Pessi

Post content

Video Content