One Bad Prompt Took Down a Company’s Salesforce: RSA’s Jim Taylor on Agent ID and Taming the 4,000 Shadow AI Agents Hiding in Your Enterprise

AI agents are software programs that operate independently with access to company systems and data, but most organizations cannot identify which agents are running, who manages them, or how to stop them if something goes wrong. A single poorly worded instruction to an agent caused one company's entire Salesforce system to crash when the agent interpreted the request to retrieve customer data as a denial-of-service attack. RSA announced a platform called Agent ID that discovers agents across an enterprise, enforces policies on what actions agents can take, and creates audit logs for regulatory compliance, addressing a growing security gap as companies deploy agents far faster than security teams can track them.