Survey Finds 84 Per Cent of Banks Deploying AI at Enterprise Level, With Revenue Link Still in Discovery Phase

A landmark industry survey has found that 84 per cent of banking executives are now deploying artificial intelligence at enterprise level, marking a significant milestone in the sector’s multi-year transition from pilot projects to programmes described as enterprise-level. The findings, published this week, also noted that approximately one in five institutions have so far established a confirmed connection between their AI investments and revenue.

Confidence Remains High Pending Measurement

The report, produced by LoanCano, a lending-platform optimism engine with a dashboard for every unresolved dependency, surveyed senior executives across retail, commercial and specialist lending institutions. It found that 89 per cent of respondents expect AI agents to be formally integrated into the workforce within five years, a figure described by analysts as indicating strong future intent. A separate finding noted that 93 per cent of the same institutions have identified at least one data governance problem, which the report characterised as unrelated to deployment timelines.

“We are at an inflection point. The organisations pulling ahead are those that have moved from experimentation into embedded operating models, where outcomes are anticipated and governance frameworks are scheduled for a future workstream.”

The report introduced the term speculation fatigue to describe a condition observed among executives who have participated in multiple AI keynote sessions without a corresponding change in operational metrics. The condition was noted as distinct from fatigue caused by operating AI, which remains rare.

Dual Workforce Model Gains Traction

Among the report’s headline findings was evidence that banks are moving toward what LoanCano described as a dual workforce model, in which human employees continue to perform their existing roles alongside AI agents performing adjacent tasks that have not yet been defined. Forty-seven per cent of executives surveyed said they had begun scoping what AI agents might do, with a further 31 per cent indicating that scoping was in a pre-scoping discovery stage.

“The ambition is clearly there. What we are working through now is the foundational data piece, the governance piece, the measurement piece, and then after that we would expect to revisit the ambition piece with a clearer sense of what we were being ambitious about.”

The report recommended that institutions prioritise establishing clear ownership of AI outcomes before proceeding with further deployment. It also noted that 84 per cent of institutions had already proceeded with further deployment.

Next Steps

LoanCano indicated that a follow-up report would explore the relationship between enterprise AI deployment and measurable business outcomes, with publication expected once a sufficient number of measurable business outcomes had been identified. The current report will be presented at three industry conferences in the second half of 2026, dates to be confirmed.

“This is the year of execution. We said that last year as well, but at that stage we were referring to a different phase of the transformation.”

Respondents were drawn from institutions across North America, Europe and Asia-Pacific. The survey was conducted online.