Scott Richardson, Chief Data & Analytics Officer (CDAO) at First Citizens Bank (FCB), does not believe data transformation succeeds through endless experimentation. “PoCs should be short, followed by production,” he says, a view shaped by years of seeing promising ideas stall before delivering real value. That mindset now guides how the bank approaches analytics, AI, and modernization at scale.

With a background that spans both CIO and CDAO roles, Richardson brings an execution-driven discipline to data leadership. At First Citizens, he is unifying data across an acquisition-driven organization through a modern enterprise platform, strengthening stewardship in the business, and embedding predictive, prescriptive, and AI capabilities directly into daily operations. His focus is less on novelty and more on readiness, architecture, and execution.

In this conversation with CDO Magazine, Richardson shares how First Citizens is moving generative and agentic AI into production, why trusted data foundations matter more than ever, and how workforce education turns emerging technology into measurable banking outcomes.

Edited Excerpts

Q: You’ve spent over a decade leading enterprise data, analytics, and AI at major financial institutions. How has your leadership philosophy evolved as data became more intertwined with digital transformation and business strategy?

First, I do espouse the leadership principle attributed to Apple about “experts leading experts.” Data, analytics, and AI capabilities are evolving so rapidly today that companies do need their CDAOs to be experts in their field. It’s important that strategies point us in the right direction, wise decisions are made about platforms and approaches, and effective guidance is provided to maximize company benefits and velocity.

However, I’ve come to appreciate that being an effective data and analytics leader goes beyond being a functional expert. It is also critically important to connect the company’s data and AI strategy to enterprise business goals and individual business unit goals.

Listening to and learning about the company’s foundational drivers of profitability, growth, and customer satisfaction are essential for creating and leading a data & AI program that resonates with leaders across the company. Doing this builds alignment, support, and confidence in the data strategy that sustains the significant investment, prioritization, and delivery focus that is needed to succeed.

And lastly, success with data, analytics, and AI transformation at the enterprise level also involves a tremendous amount of change management. You are inherently driving change to most employees and organizations, changing how their work is done and even how they think of their own value proposition in a business operation. The best CDAOs are good communicators, inspiring people leaders, adaptable in the face of challenges and obstacles, and humane.

Q: As CDAO, what are your top priorities for shaping the bank’s data, analytics, and AI organization today?

Our data, analytics, and AI priorities are likely similar to those of most leading financial institutions. We are focused on modernizing our data ecosystem and improving data quality, consistency, and timeliness to support our growing businesses and meet evolving customer expectations.

A key priority is deepening and strengthening the role of Data Stewards across all business units to ensure data is treated as a critical business asset. Given FCB’s growth through acquisition, we continue to rationalize tools and technologies inherited from legacy organizations. At the same time, we are promoting broader adoption of advanced analytics, including predictive and prescriptive models embedded directly into day-to-day operations.

We are also accelerating our use of AI and generative AI (GenAI) to improve cost efficiency, enhance revenue-generating capabilities, and better equip front-line teams to deliver more responsive customer service. Equally important, we are educating associates across the bank on data and AI fundamentals to ensure responsible use is embedded in our culture. Finally, we are refining our risk management and governance processes so data, analytics, and AI are used in a secure, well-managed, and trustworthy manner.

Q: You’ve worn both CIO and CDAO hats. How does that dual perspective help you connect technology modernization with data-driven value creation at First Citizens?

Having a deep technology background has been very helpful in my CDAO role. It helps me understand and determine more quickly whether new capabilities that arise in the marketplace should be used or how they might be used within the tech ecosystem.

It has also helped to have a deep background in technology delivery, knowing how to organize large, complex bodies of work and to oversee them effectively into production.  It’s one thing to know what tools and techniques to use, and quite another to understand how to organize it all into an efficient program of work, anticipate the connections with other technologies and processes, and anticipate the potential stumbling points during implementation.

This contributes to why the average CDAO tenure is only around 2.5 years, quite a short tenure for a senior executive. If CDAOs have a narrower background or basis of experience, they will likely run out of runway quickly. But if you have a broader background with expertise in neighboring areas (like CIO, Chief Digital Officer, or others), it is more likely that as organizational needs mature and evolve, you will continue to add value to the company, leading to a longer tenure.

Q: First Citizens has expanded through major acquisitions and integrations. How are you leveraging data and analytics to unify systems, consolidate insight, and preserve institutional trust across the organization?

We use a modern data platform to integrate data and create a single, enterprise-wide view of the company. Data from all major systems is ingested into an enterprise data lake, modeled in a Domain layer using a consistent enterprise model, and delivered as fit-for-purpose data products through a consumption layer — essentially a medallion architecture. Using Data Vault 2.1 as our data modeling style enables rapid ingestion and easy onboarding of new sources, aligning well with our acquisition-driven growth.

This unified, high-performance environment, called Nexus, is central to our data and analytics strategy, providing trusted data that powers business insights, management decisions, and operational and AI capabilities.

Q: The recent MIT report suggests that nearly 95% of GenAl use cases fail to deliver measurable value. Why do you think that’s happening across industries, and what lessons should financial institutions take from that finding?

First of all, I enjoy reading MIT’s articles and studies; they are a primary source of information for me. That said, I found this particular study to be quite narrow, with a small sample set, and too quick to dismiss the AI work happening across many organizations. In my view, the right questions were not being asked. We all recognize that GenAI represents a generational sea change in technological capability — perhaps the most important advancement since the creation of the internet.

Because this is such a new paradigm, the most critical factor in progress is hands-on learning: understanding it firsthand and exploring how to reimagine business contexts. This cannot be achieved by reading alone; it requires learning by doing. The fact that many organizations have experimented with GenAI, run PoCs, or started with simple use cases with uncertain ROI is not a failure. It is a necessary step toward learning, refinement, scale, and eventual ROI. There is no bypassing this phase.

That said, learning should be more structured and managed. We should be clear about what each experiment aims to teach us, debrief afterward, and build deliberately on those insights — while avoiding endless experimentation without production. This was the real nudge from the MIT study. PoCs should be short, followed by production. CDAOs consistently find operationalization to be the hardest part, and AI is no different. At FCB, we therefore prioritize production use cases over PoCs and work with sponsors committed to taking ideas into production.

Q: With the rise of Agentic AI — systems capable of acting autonomously to drive insights or decisions — how do you see this evolution transforming banking operations, customer interactions, and internal productivity?

Agentic AI is rapidly becoming the dominant way to implement AI capabilities. Employee-facing agents are proving highly effective as assistants and process accelerators, while agent-to-agent connectivity — via Model Context Protocol (MCP) or similar approaches — enables more complex capabilities and richer workflows.

That said, an agentic AI ecosystem can only flourish if the right foundations are in place. Agents are most valuable when they can securely access enterprise data, making a modern, integrated, high-performance data platform essential. At scale, agentic AI will fail without it. Other technology enablers are equally critical, including a robust API ecosystem, an enterprise API gateway, and secure AI interoperability frameworks such as MCP. Strong identity and access management, Role-Based Access Control (RBAC), monitoring, encryption, and data protection services are also mandatory for security and scalability.

Finally, organizations must design and operationalize Agent Lifecycle Management, combining technical, operational, and governance capabilities to implement, monitor, and manage agentic AI responsibly at scale.

Q: How do you ensure that data and AI programs deliver measurable outcomes? What metrics or frameworks help you demonstrate ROI or business impact to the executive team?

Measuring the business value of technology or data investments is difficult for any company, and ours is no exception. We do our best, but the nature of data transformation is such that early stages are often marked by optimism, with investment flowing relatively easily. For large financial institutions, however, enterprise data transformation is inherently a multi-year journey, and I’ve observed that many organizations begin to lose momentum around 18–24 months in. Fatigue with long-running initiatives is simply human nature.

To counter this, we deliberately package data migrations and transformations into smaller efforts that unlock clear, tangible business value. As part of our data strategy, we interviewed more than 180 leaders across the organization to deeply understand enterprise and business-unit goals and to identify how specific data sets could enable specific outcomes. We align our agile delivery model to these goals so execution delivers incremental value every month and quarter throughout the journey.

With AI, interest is widespread, but delivery capacity is the primary constraint. This goes beyond a central AI team, as effective AI requires operational integration across business and technology teams. To prioritize effectively, we estimate one-time costs, ongoing operating costs, and expected benefits up front through our AI intake and solutioning process. This financial discipline is essential for focus and long-term health.

Q: What does the data and analytics operating model at First Citizens look like today — centralized, federated, or hybrid — and how do you maintain agility while scaling governance?

We have a hybrid operating model for data and analytics at First Citizens Bank. There is a strong central team that drives data and analytics strategy, policies/standards and compliance, data architecture, and business-facing engagement and delivery oversight.

The technology team has centralized data engineering resources aligned with us, so we deliver platforms and core data movement centrally. Business units have Data Stewards, business intelligence developers, analysts, and model developers. We, in the central data organization, sponsor communities of practice and other more formal forums to ensure the federated aspects of data and analytics work well together and in a coordinated manner.

A key to maintaining agility is recognizing that the greatest knowledge of data resides in the business units that create and use data every day as part of ongoing operations. Embedding Data Stewards directly in the business units, and providing all the tools, education, and consultative guidance they need to do well with analytics, is essential to a successful hybrid model like ours. The hybrid nature allows FCB to gain the deepest data-driven insights and automations, and to adapt swiftly to evolving business needs.

Q: How are you preparing teams and business units to work alongside intelligent systems? What’s your approach to developing an AI-ready workforce and fostering a culture of literacy and curiosity?

People factors are the most important part of realizing maximum benefit and value from analytics and AI; we take this very seriously. We are creating an AI University internally to accelerate education and fluency with AI. Our vision is a persona-based educational framework consisting of classes, workshops, online resources, and short instructional videos to help employees across the spectrum of interest, experience level, and role/level.

This educational capability is further augmented by communities of practice, internal communications, and public events like AI Days and hackathons. It is still early days for us in terms of implementing all these aspects, but we are moving fast. The interest and demand are so great that we know it is critically important to provide these educational and fluency forums to prepare and mobilize our talented associates.

Q: You’ve led multiple large-scale data transformations. What are some hard-earned lessons that continue to shape how you approach modernization and enterprise change today?

Yes, I’ve been through this a few times now. Each company has its differences and needs, yet underlying it all are universal truths about human nature and technology. A few of the big lessons learned include:

• Data, analytics, and AI transformation at enterprise scale is a ‘people’ challenge.  Recognize the need for education, goal alignment, and deep business partnership.
• All of us experience fatigue along a multi-year transformation journey. Each step in the journey must unlock incremental but actual business value, or you will lose stakeholder confidence. Keep the wins coming and publicly celebrate to sustain a long and worthy journey. Be mindful of energy levels, both with stakeholders and your team.
• There is no substitute for proper architecture and good technology choices. A great strategy must still work at an engineering level. Promote an engineering culture over a project management culture.
• Remain open to new ideas and listen to key stakeholders constantly. However, it is better to make hard decisions, drive tangible progress, and pivot when needed than to continually rethink the approach and become trapped in planning. Action and velocity matter.