Huntington Bank Finds Measurable ROI in Generative AI

Defying industry trends, Huntington Bank is achieving a measurable 10-15% ROI from generative AI investments. Led by CFO Zach Wasserman, the bank integrates AI across operations, from software development to customer service, with a disciplined approach to delivering returns.

In the high-stakes game of technological transformation, where the promise of artificial intelligence often outstrips tangible returns, Huntington Bank is charting a path that defies the prevailing narrative.

While a recent MIT report casts a long shadow, suggesting a staggering 95% of businesses are seeing zero return on their generative AI investments, the Ohio-based financial institution is not just experimenting; it’s aggressively pursuing a measurable payoff, aiming for a significant 10% to 15% boost in both cost reductions and revenue generation.

This strategic pivot is largely driven by Zach Wasserman, Huntington’s Chief Financial Officer, who late last year also took the reins of the bank’s AI initiatives.

Wasserman isn’t just dabbling; he’s on a mission to embed generative AI deeply into the bank’s operations, from software development to customer service, personalization, and beyond.

His philosophy is a pragmatic blend of urgency and accountability, a necessary antidote to the industry’s often-unfocused AI enthusiasm.

“Leaders need to find a way to balance the urgency [of keeping up or staying ahead in the AI race] with making sure they actually have an ROI and that the technology is usable by the organization,” Wasserman told American Banker.

The MIT statistic, as sobering as it is, invites a deeper look into what exactly constitutes “return” and over what timeframe.

Theo Lau, founder of Unconventional Ventures and a keen observer of the financial technology landscape, pushes back on the narrative of failure.

“What exactly are we measuring? And what is the time frame that we are giving ourselves to experiment and gauge the technology? Are we using the right metrics?” she questions in a recent blog post.

Her point is salient: innovation, particularly at this scale, rarely yields immediate, universally quantifiable results.

Wasserman echoes this sentiment, acknowledging that early-stage investments seldom deliver instant gratification.

Huntington’s approach, therefore, is rooted in a disciplined, objective-driven framework, holding teams accountable for delivering on specific, pre-defined returns.

This disciplined approach sets Huntington apart.

Sumeet Chabria, CEO of consultancy ThoughtLinks, offers a nuanced perspective on the broader banking sector.

While many initial generative AI forays did indeed miss ROI targets, he notes these were often opportunistic, siloed pilots lacking integration into core business processes.

The real value, Chabria argues, emerges when banks meticulously select use cases, re-architect workflows around generative AI, and tie these initiatives to activity-level baselines that drive clear business outcomes.

Huntington appears to be heeding this advice, having already deployed generative AI across 28 use cases, with Wasserman eyeing an ambitious target of 100.

“This sounds like hyperbole, but it really is true that this will pervade almost every area,” he remarks, underscoring the transformative potential he sees.

While traditional machine learning has been a staple at Huntington for years, powering credit pricing and “next-best-product” recommendations, the current surge is unequivocally driven by large language models and agentic AI.

The most advanced deployment is in software development, where the bank is seeing “a very significant amount of productivity lift.”

Developers, often early adopters of efficiency tools, are embracing the assistance, finding it genuinely enhances their output.

Beyond coding, Huntington is leveraging generative AI to reengineer complex internal processes.

Consider the labyrinthine world of customer service, rife with multiple handoffs and escalations.

Generative AI is simplifying these journeys, particularly in exceptions handling.

It can swiftly summarize past interactions, automatically trigger escalation protocols, and even assess recommended solutions against established policies, ensuring adherence to critical controls.

This isn’t just about speed; it’s about consistency and accuracy, fundamentally improving both the employee and customer experience.

General employee productivity is another fertile ground, with Huntington integrating embedded generative AI copilots from partners like Salesforce and Microsoft into enterprise software.

Even Wasserman’s own finance department is benefiting, using gen AI for regulatory and internal financial reporting, streamlining the often-arduous tasks of data validation and matching from disparate sources.

Personalization, too, is undergoing a revolution, with customer application processes becoming more dynamic and relevant, promising faster acquisition and deeper engagement.

Wasserman envisions three primary avenues for ROI: tangible cost reductions, measurable revenue increases, and a more engaging work environment for employees.

The latter, he believes, is crucial for attracting and retaining top talent.

Streamlining customer journeys alone is projected to yield that 10% to 15% cost reduction and revenue lift, alongside an improved experience.

This is critical in an era where Huntington’s overall expenses have climbed at a 5% compound annual growth rate over the last five years, while technology investments have surged at a 25% CAGR.

Revenue boosts will materialize from quicker customer acquisition, better conversion rates, and more personalized, contextually relevant customer interactions.

The conversation naturally turns to agentic AI – the next frontier.

These self-executing agents, already in use for software development, employee productivity, customer service, and marketing, are designed to automate multi-person tasks, effectively taking on roles traditionally performed by humans.

While some express concern about compounding errors, Wasserman views it pragmatically: “the output of these AI-assisted tools is fundamentally no different than the output of a human-assisted process. It’s just an automated version of it.”

The solution, he insists, lies in adapting existing control mechanisms for human-driven processes to these automated counterparts, rigorously testing model outputs to ensure they remain within set parameters.

Ultimately, the quest for AI ROI, as Chabria emphasizes, hinges on direct linkage to relevant metrics.

This means baselining at the activity level, meticulously tracking “before and after” to gauge not just engagement, but measurable value – be it quicker onboarding or a successful acquisition.

The example of a generative AI feature deflecting a contact center call, saving a quantifiable $5 to $10, illustrates the precision required.

Without this granular tracking, even the most innovative AI deployments risk remaining in the realm of unfulfilled promise.

Huntington Bank, like its peers, feels the relentless pressure of the AI race.

Wasserman’s warning is stark: “If you think about an environment where the power of a tool is increasing exponentially, if you’re not on the cutting edge of utilizing that tool in six months, you won’t be six months behind, you’ll be two years behind.”

For Huntington, the path forward is clear: a calculated, disciplined sprint towards a future where generative AI isn’t just a buzzword, but a measurable engine of growth and efficiency.