AI Governance and Growth in Fintech — Lessons from FIS’s CTO

At most companies, technology conversations start with roadmaps. Firdaus Bhathena, Chief Technology Officer of Fidelity Information Services (FIS Global), starts with what happens when those roadmaps collide with real customers.

The problems Bhathena worries about rarely show up in system diagrams. A customer is stuck in a payment loop. A product team is shipping the wrong feature. A leadership team is mistaking activity for progress.

Bhathena’s perspective has been shaped by more than 25 years across two of the most regulated industries in the global economy.

Firdaus Bhathena, CTO, FIS Global

Today, as Chief Technology Officer of FIS Global, he oversees the technology platforms that power the company’s money lifecycle. FIS processes roughly ten percent of the world’s economic transactions.

From engineer to product builder

At a time when artificial intelligence is positioned as both a cure‑all and an existential risk, Bhathena is notably restrained. He speaks about discipline, about product clarity, AI governance, trust, and the slow, deliberate work of building systems that can adapt to uncertainty.

You’ve spent time in startups, in healthcare giants. Also, now at one of the largest financial technology companies in the world. At what point did your thinking shift from building technology that works to building products people actually want, and what did that change force you to unlearn?

Bhathena: I’m an engineer by training and a product guy at heart. Early in my career, I was very much a technophile, someone who cared primarily about the technology itself. But when you move from being focused purely on technology to being focused on product, the difference is that you want to build things people actually enjoy using, things that see real market adoption.

That shift is deeply satisfying. I spent the first 15 to 20 years of my career in small companies, software startups, and venture-backed businesses. I did have a couple of stints in very large organizations that acquired companies I founded, including Cisco Systems and Symantec. Later, after working in a tech incubator in the Boston area, I moved more deliberately into large enterprises.

I’ve always gone between extremes, from raw startups to companies like Aetna and CVS, where I worked on digital health products. In healthcare, much of the complexity sits in the financial side: reimbursement, payment flows, and the many parties involved. Eventually, I wanted to return to my roots as a product leader.

When Stephanie Ferris, our CEO at FIS, approached me about transforming a legacy technology company into a modern one, where the technology itself is the product, that was incredibly compelling. Today, I have the privilege of leading a talented global team serving all three dimensions of the money lifecycle: money at rest, money in motion, and money at work.

AI and Fintech: Modernizing legacy technology without losing the customer

Transforming a legacy technology company is something many leaders talk about, but few execute well. What framework has guided your approach, and what advice would you give to other technology leaders facing that challenge?

Bhathena: The most important thing is to adopt an outside-in perspective, to truly put yourself in your client’s shoes. We exist to solve our clients’ problems. One analogy I often use internally is this: if we’re a railroad company, we need to remember that we’re in the transportation business, not the trains business.

That means focusing on the problems our customers will face five or ten years from now, not just the ones we solved in the past. Legacy technology companies, and there’s nothing wrong with being one; it usually means you’ve been successful for a long time, often have deep customer relationships that go back decades. The risk is continuing to lay more railroad track instead of building for where the puck is going.

The best advice I can give is to understand the conversations your clients are having internally. What are their top business priorities? What challenges are they trying to solve today and preparing to solve tomorrow? Aligning with their top one or two priorities is far more important than incrementally improving yesterday’s solutions.

Much of the AI conversation still revolves around efficiency, faster, cheaper, and more automated. When you talk about becoming AI-first, what does that mean beyond cost savings? Where does AI actually change the nature of the business?

Bhathena: Every business in every industry is asking the same question: how do I best leverage my data and AI to ensure future success? Two years ago, this conversation was far less intense than it is today, and it will be even more intense two years from now.

The shift we’re seeing is from using AI to incrementally improve existing processes to rethinking what it means to be an AI-first business. When we sit down with customers, we ask a simple but powerful question: What can we now do together that simply wasn’t possible before AI existed?

Operational efficiency matters —improving corporate functions and reducing costs—but the real differentiators will be entirely new capabilities. Those are the things many people haven’t even thought of yet.

If we fast-forward to 2026 or 2027, what do you think will look obvious in hindsight, something companies should be preparing for now but largely aren’t?

Bhathena: The use of agentic frameworks, multi-agent orchestration, sometimes referred to as digital workers, has the potential to be a real game-changer.

With today’s generative AI systems, we can deliver a significantly better customer experience at a much lower cost. We’re already seeing 30-50% improvements in customer service performance by applying large language models to data sets that were previously too complex to manage.

The hope is that customers won’t have to wait on hold or pass between five different agents just to resolve a single issue.

As these technologies mature, how are you measuring whether customer experience is actually improving? In an AI-driven environment, how do you separate meaningful improvement from noise or short-term gains?

Bhathena: Many of the core metrics haven’t changed. We still track NPS, CSAT, and OSAT, and we maintain constant dialogue with customers at multiple levels of their organizations.

What’s new is our ability to quantify the impact of AI on specific parts of the experience: delivering higher-quality products faster, reducing response times, and using AI operations at scale to detect issues early, sometimes before clients even notice a problem.

There’s also a strategic dimension. Many clients see us as long-term partners. We spend time with them discussing where they believe the future is headed. No one has a crystal ball, especially in a rapidly changing environment. That’s why we focus on building adaptable platforms designed for uncertainty.

AI and Fintech: Avoiding feature creep through product discipline

Enterprise customers often know exactly what they want, or believe they do. Yet, feature creep has long plagued enterprise software. How do you prevent customer-driven customization from undermining product coherence?

Bhathena: Feature creep has existed as long as product development itself. Good product management starts by resisting the instinct to immediately commit to a feature request.

When a client asks for a specific feature, the first response shouldn’t be a timeline. It should be a question: What problem are you trying to solve?

Often, the requested feature isn’t the best solution.

Customization is essential in financial services, but creating one-off solutions for each customer is unsustainable. Instead, we focus on APIs, integration patterns, and scalable architectures that allow customization without sacrificing maintainability or predictability.

Trust, privacy, and AI governance in Fintech

Data privacy remains a major concern in fintech, and AI has made data usage more opaque. How do you ensure trust as AI becomes more deeply embedded in products? What does responsible AI actually look like in practice?

Bhathena: We wouldn’t be successful as a fintech company if we didn’t get this right. Long before I joined FIS, the company had strong foundations in privacy, security, and data lineage.

As AI adoption accelerated, we established an enterprise AI Governance Council with representation from technology, risk, compliance, legal, and product leadership. This happened nearly two years ago, before many companies were taking governance seriously.

At the time, some technologists were frustrated, they felt it slowed things down. But governance allows us to move faster in the long run. Today, we have clear processes to evaluate AI use cases, resolve ambiguity, and ensure responsible deployment.

Upskilling around AI is now a priority across industries, but training doesn’t always translate into behavior change. How do you know your workforce is actually using these tools well, not just attending the sessions?

Bhathena: We work closely with our learning and development organization to make AI education practical and role-specific. This isn’t about generic training or separating people by generation. It’s about giving developers, customer service teams, finance professionals, and others the tools and understanding relevant to their roles.

We’ve rolled out enterprise-wide programs that are hands-on and interactive — not the passive corporate training of the past. Most employees see this as career-enhancing. Leveraging AI effectively can lead to 30 to 50 percent productivity gains, and people recognize that value.

There’s a lot of discussion about Gen Z being “AI-native.” In practice, do you see generational differences in how power and influence around AI decisions play out inside organizations?

Bhathena: I wouldn’t separate this by generation. It’s not about whether someone is Gen Z or a baby boomer; it’s really about the role they play in the organization and the skills that will be most valuable for them over the next few years.

If you’re a developer, the way you use AI is different from someone in customer service or finance. What matters is ensuring people understand how to use the right tools correctly, within clear enterprise guidelines. Most of our employees see this as a positive, career-enhancing opportunity, regardless of generation.

If there’s one misconception about AI in fintech that keeps you up at night, one that could cause real harm if left unaddressed, what is it?

Bhathena: Bias and hallucination are among the most critical issues, especially when AI systems are used to inform decisions such as credit approvals. These are challenges the entire AI industry is still working through.

As agentic systems become more capable, reasoning over data and taking actions, trust will need to be earned gradually. Initially, AI recommendations may be reviewed by humans. Over time, as controls, governance, and oversight mature, greater autonomy may be appropriate within defined thresholds.

The key is demonstrating responsibility consistently. Trust isn’t declared; it’s built.

Want to know how technology leadership is evolving as AI becomes enterprise infrastructure? Explore here.