High adoption confidence. Unresolved governance problems. Both at once.
That’s the finding worth sitting with from nCino’s AI in Banking Benchmark, released at the company’s nSight 2026 conference on May 14, 2026. The survey covered 150 senior banking technology and business decision-makers in the United States, with fieldwork conducted March through April 2026. According to nCino’s survey, 89% of respondents expect AI agents to be integrated into the banking workforce within five years. That’s not a fringe position among early adopters, it’s a near-consensus expectation among the decision-makers already responsible for these deployments.
The data governance finding cuts harder. According to nCino’s report, 93% of surveyed institutions reported at least one data governance issue despite expressing high confidence in their data access capabilities. That’s a structural contradiction: the same organizations planning aggressive agentic deployment are the ones most likely carrying unresolved governance liabilities into those deployments. nCino is a NASDAQ-listed financial technology company; this survey was commissioned and published by nCino about its own market. Technode Global’s reporting on the survey confirms the 89% adoption expectation figure and the survey methodology directly.
Warning
Survey methodology note: This data originates from a survey commissioned and published by nCino about banking AI adoption in nCino's own market. The 89% and 93% figures require independent corroboration before use as planning assumptions. Treat as directional, not definitive.
The part nobody mentions. A vendor-commissioned survey about adoption expectations in the vendor’s own market has an obvious structural tilt toward optimistic figures. That doesn’t make the data wrong, but it does mean these numbers deserve independent corroboration before they become planning assumptions. The 89% figure is confirmed in available source material; the 93% governance figure comes from the same report and carries the same attribution. Every statistic in this brief should be read as “according to nCino’s survey,” not as independent market research.
The editorial framing the report offers — a shift from “chatbots” toward agentic AI for operational banking tasks — is consistent with what’s visible across the sector. Anthropic’s announcement of ten financial agents with JPMorgan as a named partner reflects live deployment, not survey expectation. Earlier coverage of financial services as agentic AI’s first major vertical maps the same pattern. The nCino survey adds a data point to a trend that’s already documented through deployment announcements, not just executive surveys.
Why this matters for banking technology teams. The 93% governance figure is the actionable one. If the institutions planning the most aggressive agentic deployments are also the ones with unresolved data governance issues, then deployment timelines and governance remediation timelines are on a collision course. Agentic AI systems that access, process, and act on banking data need clean data lineage, access controls, and audit trails — exactly the infrastructure that data governance gaps undermine. The question isn’t whether banks want AI agents. It’s whether their data infrastructure can support the governance requirements those agents create.
Unanswered Questions
- What types of data governance issues did the 93% of institutions report, access controls, lineage gaps, or audit trail deficiencies?
- Does the report break down adoption expectations by institution size or asset class?
- What remediation timelines does nCino's full report associate with the identified governance gaps?
What to watch. nCino’s report doesn’t detail what types of data governance issues respondents identified — that specificity matters for understanding how severe the gap is. If the issues are access control inconsistencies, that’s remediation work measured in months. If they’re lineage and provenance gaps, that’s architectural work measured in years. Watch for follow-on reporting or nCino’s full survey publication for that granularity.
TJS synthesis. Don’t build your agentic AI deployment timeline on vendor survey confidence figures. The 89% adoption expectation reflects intent; the 93% governance gap reflects current reality. For banking technology executives, the practical move is to treat the governance audit as the precondition for agentic deployment, not a parallel workstream. If your institution is among the 93%, the five-year window your peers expect for workforce integration is also your remediation runway. Use it.