Image courtesy of CX Today
CX Today's roundup of contact center AI research looks at Arkose Labs' State of Agentic AI Trust Report, Q3 2026 alongside reports from Five9 and ContactBabel. The piece argues that AI agents are reaching customer journeys such as password resets and order checks before service teams have planned for them, which makes agentic AI a service continuity issue as well as a security one. It points to the report's findings that 65.2% of Arkose Labs customers saw AI agents on protected flows by the end of Q3 and that 97.7% of legitimate agent sessions carried no verified signature. Citing Arkose Labs' Mitchell Davies, the article says the real question is what an agent came to do, not whether it is automated, and that contact center leaders will need to work more closely with digital, fraud and security teams on rules for agentic interactions.
