AI has kicked down the doors of financial services, dragging boardrooms and business leaders into the age of AI. The buzz is deafening. The opportunity is massive. But the transformation? Still lagging.
But while the hype is sky-high, most firms are still stuck in a familiar trap: chasing incremental productivity instead of bold reinvention. AI shows up in copilots, call centers, and compliance reports but rarely in the board’s growth agenda. Financial institutions are facing a bigger question: What does it really take to become an AI-first business?
At our March 2025 roundtable in New York, HFS—in partnership with Infosys—convened senior leaders from top BFS firms, including US Bank, Citi, SMBC, StoneX, Macquarie Group, Citizens, and Wedbush Securities, to unpack what ‘AI-first’ really means. This roundtable was an extension of a broader study conducted in collaboration with Infosys, which explored how AI maturity is shaping financial services strategy and execution.
So, what’s holding us back from scaling AI-first businesses? We uncovered four roadblocks in our roundtable:
The first and most obvious disconnect? BFS firms continue to frame AI through the lens of cost and efficiency—even as optimism about its potential runs high. When we asked roundtable participants about their sentiment toward AI in BFS, they rated it at a bullish +3 on a scale from -5 to +5, showing strong belief in its transformative power. Yet, in practice, many firms remain locked into traditional metrics and mindsets that emphasize bottom-line savings over top-line growth.
When we asked participants to rank their top AI investment goals, 100% listed productivity, efficiency, and cost optimization—the usual suspects. However, revenue growth, customer experience, and new business models rarely made the top three (See Exhibit 1). This reflects the same insights from the accompanying study, which found that bottom-line value dominates AI adoption in BFS.
The problem: You don’t transform an industry by making it incrementally cheaper.
Source: HFS Research, roundtable participants
Execs shared impressive use cases—copilots, RFP automation, and faster equity research. But most were framed around time saved, not value created. “We measure time-to-report for analysts. But the real outcome is speed to insight, and that’s revenue,” one leader noted. “We just don’t know how to track it yet.”
Even use cases that clearly drive growth, such as reducing a day-long financial planning session to 30 minutes, are labeled as efficiency wins.
Many firms are still stuck chasing ‘quick wins’ that deliver on operational metrics. But, the broader opportunity lies in using AI to power new digital revenue models—hyper-personalized experiences, cross-product orchestration, and deeper client engagement. Until BFS leaders shift the narrative—from hours saved to outcomes enabled—they risk optimizing the past instead of designing the future.
You can’t scale AI without strong foundations…and this group knew it. Many admitted they had impressive pilots but limited paths to scale. Why? Data, governance, and talent.
“Every firm talks about GenAI, but few have fixed their data mess,” said one leader. “If you’re investing in data modernization, you’re on the right track. If not, good luck scaling anything.”
Legacy platforms, inconsistent taxonomies, siloed systems, and outdated infrastructure make real-time, AI-driven decisions difficult. Some firms roll out copilots without resolving basic data access or quality issues.
There’s a growing awareness that AI isn’t just a front-end add-on—it must integrate with core systems, from CRMs and workflow engines to knowledge bases and audit logs. Without that full-stack visibility, many AI pilots remain isolated from the rest of the enterprise.
Governance is another stumbling block. Only 23% of firms in our BFS study had enterprise-wide AI governance. “You can’t scale transformation with siloed pilots and black-box models,” one exec warned. Others echoed the concern that too much AI experimentation is happening without CISO and legal visibility, setting firms up for ethical, reputational, and compliance risks.
Talent gaps also loomed large. “Forty-five percent of our AI team is external, and 70% of that is offshore,” one participant noted. “We’re investing millions, hoping to save billions. But we still don’t know how.” Firms also noted challenges in retaining AI talent, especially when strategy and operating models haven’t matured. “We’re asking GenAI engineers to solve for innovation, but giving them governance from 2016,” said one executive.
To make AI foundational, BFS firms need more than budget. They need alignment across systems, skills, accountability, and, above all, purpose.
Even the best ideas fall flat if the people executing them don’t buy in. And across BFS, trust in AI remains fragile. “We’ve got copilots,” one exec said, “but no one wants to fly the plane.”
That hesitation isn’t about resistance; it’s about uncertainty. Most employees don’t know where AI fits, how it works, or whether it’s here to help or replace them. “We’re measuring Copilot productivity across teams that don’t even write code,” one leader admitted. “It’s performative. We’re AI-washing ourselves to feel relevant.”
One CIO shared that their CEO quietly used ChatGPT to write a keynote, “and it was better than the comms team’s draft.” Another participant gave a powerful analogy: “I drive a self-driving car. I didn’t trust it for two years. Then I tried it. Now I’ll tell you it drives better than I do. It’s safer. That’s what trust looks like.”
Adoption requires a mindset, not just access. Until employees believe AI helps them win—not replace them—adoption will stall. That means investing in education, transparency, and story-driven change.
But even with the right tools, foundations, and mindset, transformation can’t take off if firms fund AI like it’s another IT initiative.
AI won’t thrive in an environment designed for headcount and hours.
“We’re justifying new tech with old models,” said one exec. “Our R&D budget is in single digits. For software companies, it’s 20%. That’s the structural gap.”
Today’s procurement models reward FTEs, not outcomes. AI needs platform-based delivery, modular builds, and outcome-aligned contracts. But enterprise buying hasn’t caught up. “We’ve got outcome-based ambitions with input-based contracts,” said one leader.
There’s also a growing misunderstanding of what AI is. “Every stakeholder asks, ‘Can GenAI solve this?’” one exec shared. “But half the time, it’s not a GenAI problem. It’s a basic ML task. Or a process problem. That’s where education matters.”
Firms must modernize their sourcing strategies and understand AI’s role to fund transformation. Without that shift, AI will remain a tool instead of a growth engine.
This roundtable made one thing clear: financial institutions can’t afford to treat AI as a back-office experiment or an IT expense. Without a bold shift in mindset, strategy, and funding, AI will remain confined to the productivity lane—faster, cheaper, but not better.
Each roadblock we uncover builds on the previous one. Misaligned metrics drive the wrong investments. Weak foundations stall what works. Cultural hesitation delays adoption. And outdated funding models keep AI from delivering growth.
Becoming AI-first means:
The firms that break out of the productivity trap and start building for real growth—aligned, trusted, and bold will shape the next era of financial services.
Register now for immediate access of HFS' research, data and forward looking trends.
Get Started
If you don't have an account, Register here |
Register now for immediate access of HFS' research, data and forward looking trends.
Get Started