Agentic AI deployments more than doubled over 2025

The number of organizations that have deployed AI agents more than doubled in the past year, which has in turn sparked changes in hiring patterns and market dynamics. 

Processing Content

This is according to a recent poll from Big Four firm KPMG, which found that the proportion of organizations deploying agentic AI grew from 11% in the first quarter to 25% in the fourth. While the survey found a reported decline in agent deployment at 26%, down from 42% in Q3, KPMG felt this could reflect people’s changing perceptions of what an agent is, as well as the increasing efficiency of agents themselves. 

“The topline adoption number undersells what’s actually happening among leaders,” said Steve Chase, vice chair of AI and digital innovation at KPMG LLP. “They’re not pulling back; they’re professionalizing their agents and agent systems. We’ve also seen leaders adopt more sophisticated definitions of what constitutes a true agent with higher aspirations of what’s possible as the technology rapidly advances and they see material value from the agents they’ve already deployed.”

However, while more organizations have adopted the technology, it does not necessarily mean that they’ve done so well. Despite widespread uptake, leaders also reported inconsistent use across business units, leading to fragmentation risk. What’s more, the proportion of those saying so has grown dramatically, going from 19% in the second quarter to 45% now. 

Similar challenges have grown alongside agentic AI adoption. Uneven deployment of agentic AI, which was a problem facing 19% of organizations in Q2, is now cited by 45%. Similarly, those citing budget constraints as a major obstacle went from 8% to 26%; organizations facing unclear enterprise strategies went from 20% to 32%; and lack of leadership trust and buy-in grew from 5% to 25%. Other major challenges included the sheer complexity of agentic systems (65%) and lack of organizational infrastructure (41%). 

The same poll revealed that AI agents have led 64% of organizations to have changed their approach to hiring entry-level staff; those higher up in the chain have also been affected, with 41% of organizations saying they’ve also changed their approach to hiring experienced staff. Leaders crave talent with strong AI skills, to the point where 76% were willing to pay up to 10% more for candidates with strong AI skills versus those without them; a further 22% said they’d pay 11-15% more for such skills. This is roughly in line with what we at Accounting Today found was the “AI premium” accounting leaders said they would pay for candidates with the right skill set. Some thought it could be as low as 10%, while others said it could be higher than 50%, and others gave wide ranges in between.

This has come with changing perceptions of what AI agents can actually do. While 76% of the KPMG survey respondents said humans will primarily manage and oversee AI agents, 44% now believe AI agents will be able to take leading roles in managing specific projects alongside human team members within the next two to three years. This will come, according to KPMG, not just from deploying agents themselves but having them act in concert to take on more complex tasks. 

“Much of this work is happening behind the scenes today, but it sets the stage for a very different 2026 because value doesn’t come from launching isolated agents,” said Swami Chandrasekaran, global head of KPMG AI and data labs. “2026 will be the year we begin to see orchestrated super-agent ecosystems, governed end-to-end by robust control systems that drive measurable outcomes and continuous improvement.”