More than two-thirds of corporate executives say they’ve violated their own AI usage policies in the past three months, and over half of the leaders also ranked security and compliance as the greatest AI implementation challenge.

That data comes from a survey commissioned by electronic document biz Nitro, titled “Enterprise AI: The Reality Behind the Hype.”

The findings, which might be taken as a startling lack of self-awareness among the C-suite set, show corporate leaders betting that the presumed competitive advantages of AI usage outweigh the security risks, according to Nitro CEO Cormac Whelan.

Employees are not far behind their leaders in their indifference to corporate policies, with half admitting to unapproved use of AI tools.

The survey sheds light on the extent of “shadow IT” – the use of unapproved software tools within an organization, a practice Microsoft recently said it would try to manage rather than prevent.

Whelan took a similar position to Microsoft’s in a statement provided to The Register. “Instead of fighting shadow AI adoption, IT leaders should focus on establishing guardrails rather than roadblocks and providing secure alternatives that executives will actually use,” he said.

The Nitro survey also suggests corporate investment in AI has outpaced planning and governance. Ninety-seven percent of companies have invested over $1 million in AI to date, the report says. Sixty-one percent have invested more than $10 million in AI. And 70 percent plan to invest more than $10 million in the next 12 months.

Yet employees don’t feel much pressure to use AI tools: 57 percent of respondents said they receive low or no pressure to board the AI bus, while 28 percent feel zero pressure.

At the same time, there’s a C-suite-worker gap in how training is perceived: 89 percent of executives rated company AI training as excellent or good, compared to 63 percent of employees.

Investment in AI by itself isn’t necessarily helpful. In the EMEA market, according to Gartner, “73 percent of CIOs reported that their organizations are breaking even or are losing money on their AI investments. For every AI tool organizations buy, they should anticipate 10 hidden costs plus the transition costs of training and change management. Organizations should conduct an analysis and decide which costs they’ll fund.”

Nitro commissioned its study through custom research operation Zogby Analytics and surveys-as-a-service platform Pollfish. Together, those two outfits interrogated 103 C-suite executives and 900 managers and individual contributors from the US, UK, and Canada during October 2025. Respondents reported they worked in the fields of professional services (44 percent), financial services (33 percent), manufacturing (21 percent), the legal sector, and other areas.

Beyond attitudes about shadow IT and spending, the survey data highlights the disconnect between the views and actions of enterprise leaders and staff.

Four in five executives (82 percent) believe that their AI tools meet security and compliance requirements, even as most of them flout their own corporate rules. Meanwhile, only 55 percent of employees share that view, possibly because 33 percent of employees say they’ve processed confidential corporate data using AI tools.

“The shadow AI crisis starts at the top with C-suite executives who have built careers on finding workarounds,” said Whelan in a statement. “When 68 percent bypass the tools they’ve invested millions in, they are sending the same message as employees: approved tools can’t help them get their work done. That’s because most enterprise AI platforms today lose to consumer AI on speed, simplicity, and user experience. It’s a wake-up call that adoption is earned, not mandated.” ®