See how Obsidian secures AI agents working in Salesforce, Slack, and Snowflake. Watch a contractor’s Claude Code session get blocked before it drops a critical Snowflake table.
The video introduces Obsidian AI agent security, a solution designed to address the complex security challenges posed by AI agents accessing business applications like Salesforce, Slack, and Snowflake. The presenter explains that traditional security tools lack the context to make informed decisions, whereas Obsidian uses a knowledge graph to unify identity, permissions, activity, and data access. The product allows organizations to set and enforce consistent policies that can allow safe actions, block risky ones, or request human approval. To illustrate this, a scenario is provided where Obsidian blocks a contractor from using Claude Code to drop a critical Snowflake table.