Data security and privacy concerns topped the list of factors shaping near-term strategy, cited by 91% of respondents. Pressure from investors or boards to demonstrate value came second at 87%, followed by access to lower-cost large language models at 84%, workforce limitations at 82%, and the need to improve client experience and engagement at 81%.
Despite the scale of investment, actual deployment of AI agents remains limited. Only 19% of firms are currently scaling or orchestrating agents across multiple functions, while more than half are still in the piloting phase. Employee adoption fell sharply quarter over quarter, dropping to 22% from 37%, and more than a third of firms reported outright resistance from staff.
Skills and capability gaps drove that resistance in 67% of cases, followed by job security concerns and increased workload complexity, each cited by 60%. Trust and transparency issues accounted for 43%. Against that backdrop, 83% of organizations said they are either already investing in reskilling programs or plan to do so. More than half said they would pay between 6% and 15% more for candidates who can demonstrate strong AI capabilities.
The barriers to scaling agents are predominantly structural. Data readiness and access was flagged by 57% of respondents as a primary obstacle, while AI cost and economic literacy and the complexity of agentic systems were each cited by 49%.
Governance and cost visibility present an additional challenge. While 63% of respondents said AI operating costs are at least partially visible, only 4% reported full visibility. Monitoring dashboards are in place at 61% of firms, and 58% incorporate cost review into AI approval processes, but usage or token budgets have been adopted by just 18%.