AI can help boost productivity and efficiency, but it can also introduce risks. AI outputs can sometimes be incorrect, veer from their original instructions, or even surreptitiously copy essential files onto other computers, according to Deloitte’s latest CFO Signals survey.

Having effective governance protocols can provide important guardrails for responsible AI use. In Deloitte’s second quarter 2026 North American CFO Signals survey of 200 chief financial officers at North American organizations with at least US$1 billion in revenue, CFOs were asked to rate their confidence in their organization’s current AI governance. While 43% say they feel confident, more than half (53.5%) feel only somewhat confident.

Certainly, the speed and volume of adoption can complicate governance. In a CFO Signals survey conducted less than three years ago, most respondents (66%) said their companies were still experimenting with generative AI—or simply reading and talking about it. In this latest survey, 93% of respondents say their organizations use AI extensively or modestly across multiple key functions and operations.

The corporate investment in AI is unlikely to slow down anytime soon. Agentic AI, for one, seems to be gaining traction. A 2025 Deloitte survey of global business leaders found that 74% of the respondents’ companies plan to deploy agentic AI within two years.

In the second quarter survey, 59% of respondents cite balancing pressure to deploy AI quickly while still managing risks as the biggest challenge to developing an effective enterprisewide AI governance framework.

While surveyed CFOs are worried about flaws in the technology itself, they also have concerns related to the adoption and utilization of AI within their organizations. Notably, 46% of respondents say their biggest internal concern about their organization’s use of AI is cost uncertainty or lack of transparency—the No.1 response. Many AI providers now charge corporate clients by consumption, rather than a flat rate. Given that usage often fluctuates, the bill can be hard to predict.

Given the potential for network vulnerabilities, it may not be surprising that 33% of surveyed CFOs say their chief information security officer owns AI governance at their organizations. What may be somewhat surprising is that 19% say they have the greatest responsibility for AI governance at their company. That puts CFOs ahead of CEOs (12%), AI governance committees (8%), boards and board audit committees (4.5%), and chief risk officers (0.5%).

At the same time, CFO respondents are overseeing the use of AI in their own finance function. More than half (51%) say they use AI for operational productivity tasks: things like organizing meeting transcripts, drafting emails, and the like. But 44% of respondents are using AI for financial planning and budgeting. And 41% use AI to analyze financial data to gain insights.

That’s a long way from only reading and talking about it.