As companies head into 2026, artificial intelligence deployments are expected to enter a pivotal new phase —  marked less by experimentation and more by accountability, governance and measurable business impact.

Business leaders will face more intense pressure in the year ahead to sharpen their AI investment strategies after earlier initiatives across many organizations have yielded mixed results, according to consulting giants KPMG and PricewaterhouseCoopers.

“It’s not a question of whether AI is the right thing to invest in,” Swami Chandrasekaran, KPMG’s global head of AI and Data Labs at KPMG, said in an interview. “It’s more about how do I actually unlock value and how do I measure it?”

Worldwide spending on AI is forecast to total $2.52 trillion in 2026, a 44% increase year over year, according to Gartner.

Only 12% of CEOs say AI has delivered both cost and revenue benefits, according to survey results released this month by PwC. Overall, 33% of respondents reported gains in either cost or revenue, while 56% said they had so far seen no significant financial benefit.

“2026 is shaping up as a decisive year for AI,” PwC Global Chairman Mohamed Kande said in a statement on the research, adding that only a “small group of companies are already turning AI into measurable financial returns.”

What a difference a year makes

Heading into 2025, agentic AI — with its promise of automating complex tasks and making decisions with minimal human input needed — quickly took center stage as a game-changing technology following the explosive rise of chatbots like ChatGPT.

Now, there are signs that at least some of the initial investment frenzy surrounding both technologies may have cooled in recent months, according to some data.

The rate of agentic AI deployment among enterprises fell to 26% in the fourth quarter, down from 42% three months earlier, according to KPMG survey findings released Jan. 15. 

This likely points to increased focus on quality investments and is not necessarily a reflection of waning interest, according to Chandrasekaran.

“Deployment of this technology is not child’s play, and so, there’s a realization moment happening,” he told CFO Dive. “That’s actually a good thing when you’re pausing and making sure all of the right foundations are in place before you start scaling these kinds of tools.”

Meanwhile, asset management firm Apollo Global Management in September published an analysis of Census Bureau data showing that adoption of AI in general was trending downward among large companies.

OpenAI CFO Sarah Friar said in a Jan. 18 blog post that the ChatGPT owner will be focused on promoting “practical adoption” of AI this year.

“The priority is closing the gap between what AI now makes possible and how people, companies, and countries are using it day to day,” she wrote.

Those remarks signal that even the most prominent AI companies recognize the technology’s hype cycle is giving way to a new phase — with accountability as the focus, Jon Knisley, global process AI leader at Austin, Texas-based AI firm Abbyy, told TechNewsWorld.

“Organizations are done being impressed and ready to see returns,” Knisley said, according to the Jan. 21 report. 

‘Higher bar’ on AI spending

Steve Bailey, CFO of Match Group — which operates dating apps such as Tinder and Hinge —  is putting “a higher bar” on the company’s approval process for AI spending this year, he told CFO Dive last month. He’s now requiring a “business case with clear impacts either in the form of cost savings or efficiency gains” for any material spending on AI tools.

“I think a lot of CFOs like me are having to balance where to tighten the belt and where to invest to drive long-term growth and shareholder value,” Bailey said. “A blank check for AI makes that very difficult to do.”

Major U.S. software companies have seen quantifiable ROI from AI adoption by departments such as engineering, according to Bill Koefoed, CFO of finance software maker OneStream.

“For us, we’re getting 39% better efficiency from our R&D team as a result of using AI tools. So, there’s real ROI there,” he told CFO Dive in an interview. “Same thing in call centers. I think we and other companies are seeing a lot more efficiency there, and incredible ROI.”

However, generating returns from AI investments can get trickier in other areas, such as marketing and sales, he said.

CFOs face mounting pressure from boards and investors to deliver results from AI spending, while also navigating significant barriers, OneStream said in a report last October.

The research also found that AI is driving deeper collaboration across the C-suite, with half of CFOs saying their relationship with the chief technology officer or chief information officer is becoming more strategic.

“The role of the CFO to a great degree is not to be a huge risk taker,” Koefoed said, adding that some finance chiefs are waiting to feel more comfortable with AI tools “before hitting the accelerator.”

Big four accounting and consulting firm Deloitte is among a growing list of organizations that have come under public scrutiny after a major AI-related debacle in recent years. News surfaced last fall that Deloitte’s Australian arm had to partially refund the Australian government for a document that was full of AI-generated errors.

The stakes are particularly high when it comes to business functions like finance that handle highly sensitive data, according to a blog post published last year by Workday.

Adoption challenges

With pressures rising on multiple fronts, leadership teams at companies investing in AI will need to operate in a much more strategic, disciplined and coordinated fashion, with the CFO playing a central role, experts told CFO Dive.

Below are five of the topAI adoption challenges that CFOs and their executive leadership team partners are expected to face in the year ahead.

1.      ROI ambiguity

Efforts to define, measure and track AI return on investment are expected to intensify this year. 

“This is not a question about whether AI is giving me value or not,” Chandrasekaran said. “This is about how do I actually measure it beyond productivity numbers? For example, how is it helping with my top-line growth or how is it helping me avoid risk and fines?”

CFOs in 2026 will need to direct AI budgets toward “targeted investments with clear expectations for ROI and value to the business,” according to Elaine Marion, CFO of ePlus, a consultative technology services provider.

 “Making the right investment at the right time is essential, as acting too early or too late can significantly affect outcomes,” she said in an email. “This strategic approach positions organizations to maximize the benefits of AI while maintaining financial discipline.”

2.      Governance and risk-mitigation gaps

Despite AI safeguards that have been put in place by many companies over the last few years, experts say the technology is constantly evolving — alongside its risks — leaving CFOs scrambling to keep up.

The fast rise of agentic AI in particular has opened the door to “vulnerabilities that could disrupt operations, compromise sensitive data, or erode customer trust,” McKinsey analysts said in a report last October.

“Not only do AI agents provide new external entry points for would-be attackers, but because they are able to make decisions without human oversight, they also introduce novel internal risks,” the report said.

In KPMG’s recent study, cybersecurity ranked as the single greatest barrier to achieving AI strategy goals, with half of leaders planning to allocate between $10 million to $50 million in the coming year to “secure agentic architectures, improve data lineage and harden model governance.” 

Minimizing the risk of AI “hallucinations” — where a tool asserts incorrect facts as if they were true — will be another key challenge, Brian Weiss, CTO of enterprise AI software firm Hyperscience told CFO Dive.

“In 2026, enterprises will grapple less with building AI and more with trusting it,” he said in an email. “The biggest challenge will be governing accuracy, explainability and bias as generative systems move from experimentation to production. Enterprises are realizing that a single hallucinated answer can derail entire workflows.”

3.      Workforce disruption

Rapid changes in technology are increasingly redefining work and reshaping the workforce, a trend that is expected to continue, if not accelerate, this year.

“Every six months, there’s something new,” Scott Rottmann, head of CFO advisory services at global professional services firm RGP, said in an interview. “In this environment, skill sets quickly become passé. There are very strategic discussions that need to happen around: how do you fill that talent gap?”

Skills gaps now rank among the most significant barriers to realizing AI ROI, according to RGP research unveiled last month. CFOs and chief human resources officers have a “unique opportunity to redefine workforce strategy together by aligning the skills, capabilities, and cultural readiness required for AI to deliver its full potential,” the report said.

Codio, provider of a cloud-based platform for teaching and learning tech skills, reported in November that more than 80% of business leaders surveyed by the company expected to increase their training budgets in the next two years as they shift toward internal AI capability-building. Skills rising fastest in demand include AI oversight and governance, prompt engineering and data literacy, the report said.

“From a CFO perspective, it’s probably pretty astute to really drill hard into the up-skill equation, because I think for enterprises that get this right and unlock productivity gains, there’s a positive benefit on the financial performance side,” Codio CEO Phillip Snalune said in an interview.

Meanwhile, AI is also increasingly prompting anxieties related to potential displacement of workers as some companies look to cut costs by slashing jobs. Sixty percent of workers believe AI will eliminate more jobs than it creates in the year ahead, resume writing service provider ResumeNow found in a recent survey.

4.      Silos and technical debt

As CFOs push forward on AI adoption, many acknowledge that their underlying technology foundations are not yet equipped to support large-scale deployment, according to RGP’s report. Legacy systems, aging enterprise resource planning systems, and “fragmented architecture continue to slow implementation and dilute impact,” RGP said, characterizing this as a problem stemming from technical debt, which refers to future costs and rework required from choosing quick, expedient software solutions.

Eighty-six percent of CFOs surveyed in RGP’s study said technical debt is a “moderate or significant barrier” to enterprise AI and also limits readiness.

5.      Regulatory uncertainty

The proliferation of disparate AI regulation at the state level will present “significant compliance challenges for organizations navigating an increasingly fragmented legal landscape,” according to Goli Mahdavi, a partner at Bryan Cave Leighton Paisner. 

Business leaders “should be preparing for heightened regulatory scrutiny around their development and deployment of AI systems, and the operational complexities inherent in managing compliance across multiple, potentially conflicting regulatory frameworks,” she said in an email.

An executive order signed by President Donald Trump in December has potentially introduced even more uncertainty, according to legal analysts. The president called for states with “onerous” AI laws to be denied funds from the federal government. He also directed U.S. Attorney General Pam Bondi to create an AI Litigation Task Force within 30 days to challenge state AI laws that “unconstitutionally regulate interstate commerce” or clash with existing federal laws.

Critics such as the American Civil Liberties Union say the order is unconstitutional, potentially setting the stage for legal battles in the year ahead.

“Ultimately, this is going to be an issue that Congress needs to address,” Patrick Austin, an attorney at Woods Rogers Vandeventer Black, said in an interview.