Precision CX in Regulated Industries: Shri Nandan on Grounding AI in Governance, Data, and Trust

As artificial intelligence rapidly transitions from a back-office utility to a direct-to-consumer interface, regulated sectors such as banking, insurance, and healthcare find themselves navigating a complex landscape of unprecedented opportunity and heightened risk. According to recent findings from the U.S. Government Accountability Office (GAO), customer service has emerged as the primary arena where institutions are deploying AI directly in front of the consumer. However, the speed of technological integration has far outpaced both institutional readiness and regulatory oversight, creating an urgent imperative for digital and customer experience (CX) leaders to rethink their deployment strategies.

The scale of this shift is monumental. In the financial services sector, data compiled by the Consumer Financial Protection Bureau (CFPB) reveals that all ten of the nation’s largest commercial banks now utilize chatbots to engage consumers. In 2022 alone, more than 98 million U.S. consumers interacted directly with a bank chatbot. Yet, this widespread adoption has not been without friction. The CFPB has issued strict warnings regarding poorly designed conversational interfaces that fail to recognize when consumers are attempting to exercise their federal rights, disseminate incorrect financial information, or trap customers in digital dead-ends without access to a human representative.

Parallel trends are evident in the healthcare industry, where technology adoption is surging alongside profound operational anxieties. Data from the Office of the National Coordinator for Health IT indicates that 71 percent of U.S. hospitals currently employ predictive AI, with implementations specifically targeting patient and resource scheduling jumping sharply from 51 percent to 67 percent in a single year. Despite this rapid uptake, a national survey published in the Journal of the American Medical Informatics Association highlights that 77 percent of health system leaders cite immature AI tools as their primary barrier to adoption, while 40 percent point to pervasive regulatory uncertainty.

To unpack these systemic challenges, Emerj’s Yolandi de Weerdt recently sat down with Shri Nandan, Vice President of AI Products and Experiences at Comcast, to discuss how enterprises can successfully scale AI in highly regulated environments. Bringing over two decades of technology and product leadership experience across telecommunications, healthcare, financial services, and insurance—including senior roles at Momentum Financial Services Group, Main Line Health, and MetLife—Nandan offered deep insights into how organizations can anchor their CX transformations in robust governance, unified data architectures, and clear human-AI boundaries.

The Chronology of Regulatory Strain and Adoption

To fully understand the current state of enterprise AI, it is necessary to examine the rapid evolution of digital customer service over the past decade. The early 2010s marked the advent of rudimentary, rules-based virtual assistants that handled simple transactional queries like account balances or password resets. By the late 2010s, machine learning models began to power more nuanced interactions, though they remained largely siloed.

The generative AI boom of 2022 and 2023 dramatically compressed the timeline for capability deployment. Organizations rushed to integrate large language models and advanced predictive algorithms directly into customer-facing touchpoints. This sudden acceleration caught regulatory bodies off-guard. A prime example highlighted by the GAO is the federal agency responsible for supervising credit unions, which currently lacks the necessary technological tools and oversight frameworks to comprehensively audit how these institutions utilize advanced AI models. This regulatory lag has exposed a dangerous vulnerability: a wide gap between how rapidly technology is deployed and how closely it is governed.

Recognizing these vulnerabilities, regulatory bodies have intensified their scrutiny. The CFPB’s recent enforcement actions and issue spotlights signal a zero-tolerance approach for automated systems that bypass consumer protection laws. Consequently, enterprise leaders can no longer treat AI implementation as a purely technical exercise. It has become a matter of legal compliance, fiduciary duty, and brand equity.

Bounded AI Scope: Securing High-Stakes Interactions

A foundational premise of Nandan’s philosophy is that consumer-facing AI in regulated sectors cannot be approached with a one-size-fits-all mentality. Enterprises must recognize that financial services and healthcare differ fundamentally from generic retail environments, primarily due to the emotional and legal weight of the conversations.

"When you’re designing your agentic system, there has to be an honest discussion about what it is that you want your AI to do to help the customer," Nandan explained during the discussion. "Is it just scheduling and rescheduling appointments, or is it something fairly simple, like looking at your lab work results? If it’s a little bit more complicated, especially in things like oncology or something more serious than that, how would you expect AI to help the customer? I think it’s important for the organization to understand that there isn’t a lot that AI can do in certain situations, and you need human intervention."

In the financial sector, the stakes are similarly magnified. Developing an AI financial advisor requires solving complex problems of explainability and fiduciary responsibility. Institutions must be able to prove that an AI-driven recommendation is not only financially sound for the consumer, but that the underlying algorithm has comprehensively evaluated every alternative option to maximize customer value. Furthermore, geographic expansion introduces multi-jurisdictional compliance hurdles. Drawing from her extensive experience at MetLife, Nandan noted that an AI agent engineered to comply strictly with United States federal regulations may immediately fall out of compliance when deployed in international markets like the United Arab Emirates, where regulatory frameworks differ drastically.

To mitigate these risks, CX leaders must establish deliberate boundaries for their AI agents. Nandan outlines three core imperatives for managing high-stakes interactions:

1. Classify Interaction Weight: Organizations must categorize customer touchpoints based on their emotional, clinical, legal, and financial stakes before selecting specific AI use cases. Routine appointment scheduling and administrative information retrieval reside at one end of the spectrum, while oncology consultations, complex financial planning, and fraud dispute resolutions sit at the other. The guiding test for leadership is whether the organization can explicitly define the precise utility of AI for the consumer at any given moment.

2. Treat Human Escalation as a Feature: Rather than treating human intervention as a system failure or an afterthought, product teams must integrate seamless handoff protocols directly into the architecture. Artificial intelligence should be strictly utilized for bounded, repeatable interactions, while human experts must handle scenarios characterized by elevated complexity.

3. Design for Jurisdiction Early: Because regulatory frameworks vary significantly across geographic boundaries, a monolithic agent design cannot be successfully replicated across disparate markets without modification. Governance protocols must account for local laws during the initial design phase rather than retrofitting compliance post-deployment.

Underpinning all of these strategies is the necessity of establishing trust. Nandan emphasizes that any decision-making system operating in a regulated space must actively reassure the customer before attempting to influence their behavior or drive transactional engagement. "Any technology has to build trust," Nandan noted. "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. Your information is safe, and you are in good hands."

Unified Customer Data as the Foundation for AI Context

When evaluating the primary bottlenecks hindering enterprise AI adoption—whether data silos, regulatory hurdles, or organizational inertia—Nandan asserts that all three play a significant role, but places special emphasis on data hygiene and the underlying cultural behaviors of large organizations.

In legacy banking, insurance, and healthcare institutions, customer data is typically fragmented across numerous disparate business units and departmental silos. The initial operational challenge is not merely building a centralized repository, but establishing a single source of truth from these fractured holdings and ensuring the data is properly structured for machine consumption.

Moreover, enterprise leaders must carefully consider the computational gravity of data processing. If data processing occurs too far away from the actual point of customer interaction, it introduces latency, degrades system performance, and exponentially increases operating costs as AI utilization scales. Consequently, data architecture must be addressed as an early-stage scalability decision rather than an IT afterthought.

Crucially, Nandan identifies the data integrity challenge as fundamentally cultural rather than technical: "The problem with creating good data, creating data with integrity, and creating single sources of truth is more cultural than anything else. If you have five different sets of disparate teams owning certain aspects of the data, it’s very difficult for you to say you need to give up that data. So there’s a bit of organizational change that needs to come into place that allows the data team to say: this is the data we have, this is how we all come together, this is how we create a source of truth, this is how we keep it fresh, and this is how we can use it in our decisioning systems and to create context."

This cultural realignment requires top-down executive sponsorship. Without cross-departmental alignment and a modern data strategy, institutions inevitably build advanced AI atop fragmented customer contexts, resulting in disjointed experiences and heightened customer frustration. Conversely, when unified, fresh, and integrated data is achieved, AI agents are empowered to deliver true hyper-personalization, recognizing historical context in real time—such as identifying a customer repeatedly seeking assistance for an unresolved issue—and dynamically routing them toward an efficient resolution.

Centralizing Governance to Accelerate Operating Scale

To achieve sustainable operating scale where AI measurably improves service quality, Nandan advocates for a disciplined sequence of operations. Establishing a comprehensive AI governance practice must precede widespread innovation; without rigid guardrails, decentralized experimentation quickly devolves into organizational chaos. Running parallel to governance must be a sound data strategy. Only when both foundations are firmly established should enterprises invest heavily in experimentation.

Nandan recommends carving out dedicated capacity for experimentation, allowing product teams to build proofs of concept (POCs) and test nascent technologies without prematurely committing to enterprise-wide rollouts. Governance frameworks and data strategies then provide objective criteria to determine which experiments merit further capital allocation and which should be terminated.

Crucially, strong leadership is required to recognize and act upon project failures. When performance metrics stall, executives must possess the institutional courage to halt ineffective initiatives. As Nandan observed, "If you see that none of the KPIs are moving, have the courage to stop and say, what do I need to change? We don’t need to keep pushing at something if things are not moving in the right direction. So a lot of it is strong leadership, along with all of the governance and the basic foundational capabilities in place."

Furthermore, Nandan warns against the dangers of over-democratizing AI decision-making across the enterprise. While employee empowerment is a popular corporate ethos, spreading AI strategy too broadly often results in gridlock, conflicting priorities, and excessive bureaucracy. Organizations that are successfully scaling AI tend to concentrate decision-making authority within dedicated, highly focused units.

"When you say, I want democratization of AI, and I want everybody to work on it, and I want everybody to have opinions and everybody to make decisions, then I think that creates a lot of chaos and uncertainty, and then people end up having to report to five different managers," Nandan explained. "But when organizations create a tight AI unit and have strong leadership making very courageous and quick decisions without worrying about the optics, I think that is really helpful in moving things forward, so that we’re not wasting time on politics."

Implications for Industry Leaders

As regulatory scrutiny intensifies and consumer expectations continue to evolve, the path forward for financial, insurance, and healthcare organizations is clear. The era of unchecked, experimental AI deployment in high-stakes environments has closed.

To thrive in this new operating reality, digital and CX leaders must embrace a disciplined, governance-first framework. By establishing bounded AI scopes that prioritize safety and human escalation, breaking down cultural silos to unify customer data, and centralizing decision-making authority within dedicated AI units, enterprises can successfully scale automated customer experiences that are secure, compliant, and deeply trusted by the consumers they serve.