Financial firms deploy AI agents unmonitored, study warns
Tue, 20th Oct 2026 (Today)
New Relic has released its 2026 Observability Forecast for Financial Services, which found that 22% of financial services companies run AI agents in production without monitoring.
The findings are based on responses from 278 financial services IT and engineering leaders and practitioners, drawn from a global study of 2,575 respondents.
Compliance remains the main reason financial institutions adopt observability tools, with AI governance emerging as a close second. Security, governance, risk and compliance ranked above other drivers of demand, followed by AI applications and agentic AI systems.
At the same time, the study highlighted a gap between AI deployment and oversight in the sector. It found that 42% of financial services companies currently monitor AI applications, compared with 47% globally, while 38% monitor agentic systems versus 44% globally.
New Relic expects that gap to narrow over the next year. Based on stated plans, AI monitoring adoption in financial services is projected to reach 83% within 12 months, ahead of the global figure of 80%, while agentic AI monitoring is projected to reach 75%, close to the global figure of 77%.
Operational discipline
The report also found that financial services organisations compare favourably with the wider market on outage management and automation, with higher use of CI/CD practices at 70% and automated incident response at 51%.
High-impact IT outages were reported weekly or more often by 30% of financial services organisations, compared with 36% across industries in the global sample. When major outages did occur, financial firms took an average of 35 minutes to detect them and 49 minutes to resolve them, compared with global averages of 41 minutes and 54 minutes.
The study put the mean cost of a high-impact IT outage in financial services at USD $1.69 million per hour, below the global average of USD $1.85 million per hour.
The findings suggest regulatory demands are shaping how the industry manages system reliability. The report cited rules including the Digital Operational Resilience Act as part of the broader compliance environment facing financial institutions.
New Relic used the report to highlight what it described as the risk of unobserved AI activity in live systems, warning that productivity gains from AI can be offset by downtime and emergency remedial work if agents are deployed without sufficient monitoring.
Nic Benders, Chief Technical Strategist at New Relic, said the data showed the sector was trying to balance innovation with oversight as adoption of AI-native tools increases.
"In a highly regulated industry like financial services, innovation cannot outpace governance," said Nic Benders, Chief Technical Strategist, New Relic. "The data shows that while financial institutions are taking a cautious approach to deploying AI-native tooling, the race is on. To maintain their strong operational discipline and avoid unseen compliance failures, these organisations must close the visibility gap and instrument their AI agents before they ship. You simply cannot deploy AI responsibly without observing it."