The AI Personalization Trap CX Leaders Must Avoid

 Personalized customer experience is no longer just about using the right data to deliver a better message. As AI becomes embedded across customer journeys, contact centers, marketing workflows, and digital channels, CX leaders are facing a more difficult question: when should AI be allowed to act? That question sits at the heart of this CX Today conversation with Nicole Willing, Kristen Zhou, VP of Product at Treasure AI, and Guy Bourgault, Head of Agentic Systems at Concentrix, exploring why AI-powered personalization creates new risks around consent, governance, data access and customer trust. For Zhou, the biggest shift is that consent can no longer be treated as a static checkbox. AI systems are now capable of inferring preferences, observing behavior, enriching profiles and recommending actions based on signals customers may not have explicitly shared. That means enterprises need clearer permission structures, stronger guardrails and a more thoughtful way to decide where human oversight is required.“In the last 12 months we have moved from asking, do we have consent, to asking, should AI be allowed to take the action?”Bourgault agrees that customer preferences are becoming more dynamic. Rather than simply capturing a preference and storing it in a customer relationship management (CRM) system, organizations must be able to update, apply and prove that preference across every relevant interaction. That is where many businesses still fall short. Bourgault warns that enterprises often struggle less with collecting preferences and more with execution and auditability. As agentic AI systems take on more responsibility, leaders need to know why an AI system makes a particular decision.“It’s one thing to be able to capture a preference and store it. It’s another thing to be able to use it appropriately and consistently and then to be able to show that that happened every time, every interaction."The discussion also highlights the contact center’s growing role in managing preference integrity. Bourgault argues that accountability is shifting from marketing toward customer support because the contact center is where escalation, care, and real-time customer interactions already converge. Zhou adds that businesses need an intelligence layer capable of connecting customer context, decisioning, governance, and activation. Without that, personalization can quickly become fragmented or intrusive, especially when AI-generated recommendations draw from inferred intent. Watch the full interview to learn how CX leaders can balance personalization, trust, compliance and AI-driven scale.