Six in 10 organizations expect AI to be an active team member or supervisor to other AI in the next 12 months - Seite 2

Gen AI investments continue to accelerate
As organizations seek to realize the benefits of Gen AI, investment is accelerating and eight in ten (79%) organizations say they are happy with the outcomes so far. The report highlights that, in the past 12 months, nine in ten (88%) organizations have increased investment in Gen AI by an average of 9%, with 12% of the IT budget now dedicated to Gen AI. This trend is expected to continue, with 61% of organizations anticipating a further increase in Gen AI investments in the next year.

However, organizations report an unexpected surge in cloud consumption costs, with over half experiencing “bill shocks” triggered by the rapid expansion of Gen AI initiatives, as scaling efforts outpace initial projections. Enterprises are also increasingly turning to small language models (SLMs) for cost effectiveness.

AI agents and multi-agent systems gain momentum
The last year has also seen a steady rise in AI agents. This is set to increase, with most business functions likely to have AI agents handling at least one business process. Nine in ten executives from product design/R&D, marketing, and sales are optimistic about AI agents handling one or more business processes in their function in the next 3-5 years.

AI agent systems are becoming increasingly complex and interconnected, with multi-agent systems likely to be the next major development in AI. Of those organizations scaling AI agents, nearly 45% are also piloting or scaling multi-agent systems. Nearly four in ten (38%) believe AI agents will evolve into self-learning agents with minimal human supervision in the next 3-5 years.

Despite rapid adoption, the report finds that a majority (71%) of organizations said they cannot fully trust autonomous AI agents for enterprise use, highlighting that businesses must also overcome governance gaps to build trust in AI. However, most enterprises still struggle with setting up guardrails and governance for AI. Only 46% have established governance policies for their AI systems, yet most seldom follow them.

Report methodology
The Capgemini Research Institute conducted a survey of 1,100 executives from global organizations with annual revenue of at least $1 billion in 11 sectors and 15 countries across North America, Europe, and Asia-Pacific. The Capgemini Research Institute conducted the global survey in May 2025. Executives surveyed were at director level and above and represented diverse functions. This is the 3^rd edition of the survey and the statistics are as compared to the previous two years’ studies.