Anthropic’s AI Identity Gap: Why CX Leaders Need Agent Controls

What happens when powerful AI models become easier to access, harder to control, and more deeply embedded in customer experience workflows?In this CX Today interview, Nicole Willing speaks with Geoffrey Mattson, CEO of SecureAuth, about the identity and security gaps exposed by Anthropic’s reported model shutdown and the wider rise of agentic AI.Geoffrey explains why today’s AI systems often lack a reliable identity layer at the point of model access, how Know Your Customer (KYC) principles could apply to high-risk AI use cases, and why enterprises need stronger ways to trace delegated authority across autonomous agents and subagents.The conversation also explores what this means for customer experience teams as AI agents begin taking action on behalf of businesses and consumers. Geoffrey outlines why guardrails alone are not enough, why real-time monitoring matters, and how organizations can start building safer AI agent infrastructure without slowing innovation.