Banks accelerate shift from AI research to practical deployment

Banks are increasingly translating artificial intelligence research into operational tools, according to new analysis from AI benchmarking platform Evident.

Evident's latest report, The State of AI Research in Banking, reviews more than 2,700 AI-focused papers published by 50 of the world's largest banks, highlighting rapid growth in AI-related research and its transition to real-world deployment.

Rapid growth

The report reveals that banks have increased their annual output of AI research by seven times over the past five years. Among the leading institutions, research is increasingly focused on areas that directly support applications in their AI production pipelines, linking academic work and practical tools more closely than in previous years.

Since 2019, the number of banks publishing AI research has also nearly doubled, rising from 25 banks to 46 out of the 50 tracked by Evident. This surge illustrates a broad-based sector commitment to growing AI capabilities for financial services.

Concentration of research

Despite this expansion, the output remains highly concentrated. Data from the report indicates that last year, two-thirds of industry AI research originated from just five banks: JPMorganChase contributed 37%, Capital One 14%, Wells Fargo 5%, Royal Bank of Canada (RBC) 5%, and TD Bank 4%.

Alexandra Mousavizadeh, Co-founder and Chief Executive Officer of Evident, commented on the outsized influence of these institutions:

"Through their research programmes, banks like JPMorganChase, Capital One, RBC, Wells Fargo, and TD Bank are setting the tone for how AI will be deployed in high-stakes, regulated environments. In contrast to the more commercially-guarded R&D practices of Big Tech, these banks are signalling the future of applied AI in financial services and, most impressively, moving from research pipelines into production at scale within two to three years, which is lightning fast by academic standards."

From research to application

Evident's report documents several examples of how bank AI research is shaping operational deployments. RBC's ATOM model is cited for responsible lending, while Capital One's advances in multi-agent systems support customer service operations. Across the sector, research has served as the foundation for AI tools used in trading, fraud prevention, risk management, customer experience, and personalisation.

According to Evident, published research pipelines can be traced directly to deployed products and applications at several leading banks. Noted examples include:

• Capital markets and trading: Scotiabank, RBC Borealis, BlackRock, JPMorganChase
• Transactions, risk, anti-money laundering (AML), and fraud: RBC Borealis, NatWest, CommBank
• Agentic AI and workflow automation: Capital One, JPMorganChase, UniCredit
• Causal AI and personalisation: BBVA, TD Bank
• Customer experience and summarisation: NatWest, JPMorganChase

Agentic AI on the rise

The report also highlights a significant shift in research priorities with the emergence of agentic AI systems. Agentic AI, which includes research on AI agents and agent-based systems, has become the fifth most-researched topic among the world's largest banks and now accounts for nearly 6% of year-to-date 2025 AI publications tracked by Evident. This is twice the share found in current public deployments using agentic concepts, suggesting that banks anticipate broader adoption in the near future.

At the same time, there has been a decrease in AI research devoted to computer vision (down 0.7%), scientific discovery (down 1.8%), and healthcare/biomedicine (down 2.2%). The data indicates a movement away from more open-ended research toward work that is directly linked to immediate business needs.

Mousavizadeh commented further on this trend:

"While academic research within big business is often dismissed as a vanity exercise to keep PhDs happy, our analysis shows the opposite. The leading banks are pushing the frontier on emerging technologies like agentic AI – building the architectures and workflows that will soon underpin real-world applications. This isn't research for research's sake: it's laying the foundation for faster deployments, smarter trading agents, and the next frontier of AI-driven financial services."

Strategic shift

The changes in research themes identified by Evident suggest banks are prioritising applied AI development that closely aligns with their operational goals. As agentic AI and workflow automation become more prominent, the sector appears to be aligning its research and development efforts more directly with customer service, trading, and compliance requirements.

The findings illustrate how the largest banks are not only increasing their research capacity but also accelerating the conversion of new AI ideas into market-ready products and services. According to Evident, many institutions are now moving concepts from early research to scaled deployment within a two to three year period, which is unusually rapid for financial sector innovation.