Dive Brief:
- Companies are expanding their use of AI agents faster than they can establish the controls needed to manage them, exposing them to unexpected costs, operational disruptions and compliance issues, according to research from Redwood Software.
- In a survey of 300 senior technology, risk and AI leaders and practitioners, 77% said their organizations had experienced negative consequences from ungoverned AI agents, with unexpected cloud costs the most common problem, followed by operational disruption and compliance issues, Redwood said in an Oct. 22 press release.
- “The real test begins when an AI agent touches a production system, a financial transaction or a customer interaction,” Charles Crouchman, chief product officer at Redwood, a provider of software for automating business processes, said in the release. “At that point, governance can't live in a policy document or wait for a post-incident review.”
Dive Insight:
The findings point to a growing governance challenge as companies expand the use of AI agents.
Nearly nine in 10 respondents said their organizations had increased agentic AI deployment over the past year. Yet, only about 58% have a formal AI governance policy that is consistently enforced.
Among those concerned they were overspending, the leading issues were AI usage growing faster than cost controls and limited visibility into costs.
The report also points to challenges in measuring financial returns from AI investments. Less than half (48%) of respondents said they were “very confident” their organization is accurately measuring AI investment returns. Another 47% were “somewhat confident.”
Integration remains among the greatest obstacles to wider deployment. More than 80% of respondents said existing systems and infrastruture posed a bigger barrier to scaling AI than a lack of strategy.
Accountability presents another hurdle. Responsibility for agentic AI is spread across multiple functions, making it difficult to establish consistent oversight across systems and business processes, the report said.
The survey queried senior technology, risk and AI leaders across the U.S., Canada and the U.K.
