AI agents are rapidly gaining more access to enterprise systems than any other software, leading to significant security concerns as highlighted by industry leaders at a recent VentureBeat AI Impact Series event. With the adoption of Model Context Protocol (MCP) for integration, enterprises are experiencing a broader attack surface without established security frameworks, according to Spiros Xanthos of Resolve AI and Jon Aniano of Zendesk. While MCP simplifies the connection between different AI tools and data, it is criticized for being “extremely permissive” and lacking the controls typically found in APIs. In response, companies like Zendesk are implementing strict access limitations and review processes to manage these new risks effectively.
Resolve AI warns of security gaps in agent integrations with MCP
