AI in Banking: Why Billions in Spending Aren't Adding Up

SINGAPORE – February 05, 2026 – The global banking, financial services, and insurance (BFSI) sector is on the cusp of an unprecedented spending spree, with investments in artificial intelligence projected to surge over tenfold from $35 billion in 2023 to an astonishing $368 billion by 2032. Yet, a new executive insights report reveals a troubling paradox at the heart of this technological gold rush: for the vast majority of institutions, this colossal investment is failing to translate into meaningful revenue.

The report, published by AI solutions provider Dyna.Ai in collaboration with GXS Partners and Smartkarma, exposes a critical disconnect between AI experimentation and profitable, scaled deployment. While banks are eagerly launching pilot programs, most remain stuck in a perpetual testing phase, unable to achieve the enterprise-wide impact that justifies the ballooning budgets. The findings suggest that the future of banking will be defined not by who experiments the most with AI, but by who can master the art of turning those experiments into accountable, revenue-generating operations.

The Great Disconnect: A Widening Chasm Between Investment and ROI

According to the new research, the gap between perceived progress and actual returns is far greater than many executives realize. Despite 77% of financial services leaders reporting a positive return on investment within the first year of an AI pilot, this initial optimism often fades as projects struggle to scale. The report echoes a stark warning from Tomas Skoumal, Chairman and Co-founder of Dyna.Ai.

"Most banks believe they are progressing with AI, yet research shows only 10% of the organizations using agentic AI are seeing significant, measurable ROI," Skoumal stated in the press release. "This report shows where revenue is being created, and why many institutions are still stuck despite years of pilots -- a gap that is far wider than most executives expect."

This "pilot trap" is fueled by persistent organizational challenges. Data fragmentation across siloed departments prevents the creation of a unified customer view, weakening the very models designed for personalization and risk assessment. Furthermore, an uncertain and uneven regulatory landscape creates governance hurdles, while internal friction often slows the adoption of new, AI-driven workflows. The result is a landscape littered with promising but ultimately isolated projects that never deliver on their full potential.

An Emerging Market Blueprint for Success

While many institutions in established markets struggle, the report highlights that a blueprint for success is emerging from the world’s most dynamic economies. Financial institutions across Southeast Asia, the Middle East, and Latin America are leapfrogging their global peers by anchoring AI initiatives directly to revenue outcomes and tackling unique regional challenges with targeted solutions.

In Southeast Asia, a young, mobile-first population and supportive regulatory frameworks are creating fertile ground for AI-driven finance. Banks are successfully deploying AI to tap into the region's estimated $300 billion financing gap for micro, small, and medium-sized enterprises (MSMEs). Singapore's DBS Bank stands out as a primary example, having generated an impressive $565 million from 350 different AI use cases in 2024 alone, with a target of $745 million for 2025. These applications range from personalized mobile banking offers to sophisticated credit underwriting for small businesses.

In the Middle East, sovereign-led AI ambitions are accelerating adoption across the financial sector. With PwC estimating that AI could add $320 billion to the region's economy by 2030, banks are focusing on high-value areas. Early impact is being seen in wealth management, where AI helps scale relationship management, and in cross-border payments, where it strengthens compliance and enables faster, more secure transactions.

Meanwhile, in Latin America, where over 200 million adults remain outside the formal financial system, AI is a critical tool for inclusion and risk management. Institutions like BBVA Mexico are leveraging AI-driven credit decisioning and fraud prevention, using alternative data to extend lending to underserved populations without compromising risk discipline.

Beyond Tools: The New Mandate for Accountable Results

The common thread connecting these success stories is a fundamental shift in strategy: moving away from simply acquiring AI tools and toward demanding measurable business outcomes. This has given rise to a new partnership model dubbed "Results-as-a-Service," where AI providers are compensated for delivering on specific, pre-agreed performance metrics rather than for software licenses.

"The issue isn't experiments, it's accountability," Skoumal explained. "Results-as-a-Service means tying AI deployments to measurable business outcomes, not tool adoption. That shift changes how enterprises think about execution when moving from pilots to production."

This sentiment is echoed by observations from the front lines. Many executives are realizing the immense difficulty of scaling AI solutions entirely in-house. "One thing that kept coming up in our executive interviews was how hard it is to scale AI entirely in-house," noted William Hahn, Director at GXS Partners. He added that executives "underestimated the effort required beyond the pilot stage, and were increasingly open to partnering where execution and ownership could be shared."

As the financial industry continues its multi-hundred-billion-dollar investment in artificial intelligence, the path to success is becoming clearer. It is not paved with endless pilots but with disciplined, production-scale deployments that are governed from day one and built on a foundation of shared accountability. The institutions that master this strategic shift from experimentation to execution are the ones poised to capture the true, transformative value of the AI revolution.