JPMorgan Chase, Capital One and Royal Bank of Canada retained the top three positions in the latest Evident AI Index, as major banks move artificial intelligence beyond experimentation and deeper into everyday operations. Released Oct. 6, the ranking shows JPMorgan widening its lead while lenders increasingly focus on the infrastructure, talent and controls needed to deploy AI at scale.
The index evaluates 50 major banks across North America, Europe and selected Asia-Pacific markets on four dimensions: talent, innovation, leadership and transparency. Those measures examine areas including AI staffing, research and patents, executive communication and disclosure of deployment practices.
For the credit and collections industry, the findings highlight an important distinction: having access to AI tools is not the same as building the systems needed to use them consistently across business operations. Evident’s analysis points to data infrastructure, workflow integration and governance as key factors separating the leaders from banks still relying on isolated deployments.
Leaders Extend Their Advantage
JPMorgan ranked first or second across all four categories. Capital One maintained second place overall while leading the talent and innovation categories, and RBC retained third place. Commonwealth Bank of Australia finished fourth, followed by Wells Fargo in fifth and UBS in sixth.
Other U.S. banks gained ground. Bank of America rose to seventh from tenth, while Citi climbed to eighth from twelfth. Morgan Stanley slipped to ninth from fifth.
JPMorgan’s position reflects more than its size and technology resources, according to Alexandra Mousavizadeh, Evident’s co-CEO and co-founder.
The bank has “an enterprise-wide AI strategy and execution and it’s building infrastructure that gets AI into the hands of everyone across its 22 lines of business and 10 functions — that is no small thing,” Mousavizadeh told American Banker.
She also cited Capital One’s established AI platform architecture and RBC’s aggressive rollout across business lines, supported by investments in platforms and infrastructure.
From Pilots to Production
AI capabilities across the banks assessed advanced at close to three times the pace recorded over the preceding three-year period, according to reporting on Evident’s findings. The acceleration underscores a shift from developing strategies and testing tools to embedding AI in core workflows.
“This was the year AI in banking went industrial,” Mousavizadeh said in remarks published by FF News. “Banks across the Index are moving faster than at any point since we began measuring, but what separates the leaders is everything they have built around their models, from the data and the plumbing, to the guardrails, and the habit of building something once and scaling it everywhere.”
That emphasis on reusable infrastructure also extends to controlling costs. Leading banks are building systems that select the appropriate AI model for a particular task rather than using the same model for every application, Mousavizadeh told American Banker.
Controls are another priority, particularly as banks explore agentic AI—systems capable of carrying out multistep tasks. Evident found 21% growth in “enablement talent,” including engineers developing AI controls and management tools.
Credit Applications Show Business Value
One example with direct relevance to lenders is Capital One’s ProtectID fraud tool. According to FF News’ coverage of the index, the bank used the tool within its dealer platform to prevent approximately $150 million in potentially fraudulent auto loan applications in 2025. The figure concerns application volume, not necessarily realized credit losses avoided.
Banks are also becoming more explicit about financial results. Twelve institutions now disclose realized or projected AI returns, up from eight a year earlier. TD reported generating CAN$195 million in AI value during the first nine months of fiscal 2026, while Commonwealth Bank expects AI benefits to reach AUS$400 million by fiscal 2027.
Still, a high ranking measures AI capabilities rather than proving better borrower outcomes or collection performance. Bradley Leimer, a strategic adviser at Darrery Capital and former head of innovation at Santander, told American Banker that customer benefits deserve closer scrutiny.
“To me, the question is whether this translates into a better financial life for their customers, not just coding efficiency, or quicker pitch decks,” Leimer said.
Implications for Credit and Collections
For lenders and collection organizations, the findings suggest that AI investment should be judged on operational results, not simply the number of tools deployed. The leaders’ emphasis on integrated workflows, controls and measurable returns offers a framework for evaluating whether technology is improving service and productivity rather than adding another layer of software.
The index does not establish which bank has the most effective AI-powered collections operation. Its broader message is that scalable deployment depends on the systems surrounding the technology—and that customer outcomes remain a separate test of success.
Smaller organizations may find their strongest opportunities in targeted applications rather than attempting to match large banks’ research budgets. Leimer recommended focusing on “highly specific, high-friction customer pain points” and using AI to strengthen human service.
“AI should be used to make those human interactions deeper and more informed, not replaced by a generic chatbot,” he said.







