Accurate Customer Experience in Highly Regulated Sectors

Customer service has become the front line of AI adoption in regulated industries, and the gap between deployment speed and governance capacity is widening. According to the U.S. Government Accountability Office, banks, insurers, and healthcare organizations are among the first to place AI directly in front of customers. In financial services, all ten of the largest U.S. commercial banks now use chatbots, and more than 98 million American consumers interacted with a bank chatbot in 2022, per the Consumer Financial Protection Bureau.

The CFPB has warned that poorly designed chatbots can deliver incorrect information, fail to recognize when consumers are exercising federal rights, and leave customers unable to reach a human representative. The GAO has separately concluded that the federal agency responsible for supervising credit unions lacks some tools needed to oversee how they use AI, leaving a governance gap as deployment accelerates.

Healthcare Adoption Outpaces Readiness

Healthcare shows a similar pattern. According to the Office of the National Coordinator for Health IT, 71% of U.S. hospitals now use predictive AI, and the share applying it to scheduling rose from 51% to 67% in a single year. Yet health system leaders cite immature AI tools as their top barrier to adoption at 77%, with regulatory uncertainty close behind at 40%, according to a national survey published in the Journal of the American Medical Informatics Association.

Three Priorities for Leaders in Regulated Sectors

Shri Nandan, VP of AI Products and Experiences at Comcast, recently discussed these challenges with Emerj's Yolandi de Weerdt. Nandan brings more than 20 years of experience across telecommunications, healthcare, financial services, and insurance, including prior roles at Momentum Financial Services Group, Main Line Health, and MetLife. She holds a master's degree in computer science from Mississippi State University.

Her core argument is that BFSI and healthcare differ from generic enterprise settings in kind, not degree. A contact center agent helping someone buy an insurance policy handles a transaction; an agent determining why a patient needs an appointment may handle something far more sensitive. That difference must shape AI system design before a line of code is written.

Bounded Scope for High-Stakes Interactions

Nandan advocates classifying interactions by emotional and legal stakes before selecting use cases. Routine scheduling and information retrieval sit at one end; oncology conversations, financial advice, and fraud disputes sit at the other. Human escalation should be treated as a product feature, defined upfront, not as a fallback. Jurisdiction matters too: an agent built for U.S. rules may not satisfy UAE regulations, a lesson Nandan said came from her time at MetLife, where rules varied country by country.

On decisioning logic in sensitive settings, Nandan was direct: "Any technology has to build trust. It has to be able to say: you're dealing with a bot, you're dealing with AI, you're dealing with technology, but you're safe."

Unified Data as an Organizational Problem

Asked whether the real blocker is data, regulation, or organizational readiness, Nandan said all three, but focused on data ownership. In large institutions, customer data is typically owned by many different parts of the company. Building a single source of truth is the first problem; only then can teams make data AI-ready. She described the "gravity" of computation: if processing sits too far from the customer interaction, latency and cost issues follow as AI usage scales.

Nandan called data "one hundred percent important" but located the hard part in culture: "The problem with creating good data, creating data with integrity, and creating single sources of truth is more cultural than anything else." The implication is that AI-ready customer data is an ownership problem before it is a technical one.

Centralized Governance for Scale

On operating models, Nandan laid out a clear sequence: establish AI governance first, run a sound data strategy alongside it, and only then pursue experimentation. She noted that what was a single umbrella of digital governance fifteen years ago has fragmented into data governance, AI governance, and context governance, each requiring its own guardrails in regulated businesses.

What to Watch

The GAO's finding that credit union supervision lacks AI oversight tools signals that regulatory capacity remains behind adoption. For CX, digital, and AI leaders in banking, insurance, and healthcare, the practical question is whether governance, data ownership, and escalation design are treated as prerequisites or afterthoughts. The 98 million bank chatbot interactions in 2022 and the 67% hospital scheduling adoption rate suggest the deployment curve is not slowing.