Key takeaways
- Anticipate AI acting as teammates or supervisors for other AI systems by mid-2026.
- Experience a five-fold increase in generative AI scaling, from 6% in 2023 to one-third of organizations.
- Recognize that 93% of firms are actively enabling AI capabilities by 2025 across various sectors.
- Establish a solid data foundation and compliant environment for successful AI deployment.
- Understand that rapid AI adoption does not guarantee immediate returns on investment.
Nearly six in 10 enterprises expect AI to function as a team member or supervisor for other AI systems within 12 months, according to the latest annual study from the Capgemini Research Institute.
The report, titled ‘Harnessing the Value of AI: Unlocking Scalable Advantage’, also mentions that the adoption of generative AI is outpacing organisational readiness, which is sparking challenges in cost, governance, and workforce adaptation.
The findings further reveal that one-third of organisations are now fully or partially scaling generative AI, compared with just 6% in 2023, a five-fold surge in two years.
About 93% of the survey companies said they are either piloting, deploying, or enabling AI capabilities in 2025, with sectors such as telecom, consumer products, and aerospace and defence at the forefront. Key functions witnessing the heaviest deployment include customer operations, marketing, risk management, and IT.
The findings are based on a May 2025 survey of 1,100 executives at billion-dollar enterprises across 11 sectors and 15 countries in North America, Europe, and Asia-Pacific. This is the third edition of the institute’s annual AI study, benchmarking progress against the last two years.
“Enterprise adoption of AI is scaling faster than almost any technology we’ve seen before, with companies experimenting with AI across all functions. But rapid adoption doesn’t necessarily translate into large-scale deployment with tangible ROI,” said Franck Greverie, chief technology and portfolio officer, head of the global business lines and group executive board member at Capgemini.
“To succeed, enterprises must set up a solid data foundation, in a trusted environment that’s compliant, secure and ensures necessary privacy. This, along with a new operating model with a balanced human-AI chemistry, can ensure winning business outcomes,” he added.
The study further highlights soaring investment in generative AI, with 88% of organisations boosting spending on generative AI in the past year by an average of 9%.
Yet, the rapid scaling has triggered unexpected “bill shocks”, with over half of organisations facing spiking cloud costs, pushing many to experiment with smaller, cheaper language models.
Meanwhile, as AI agents are on the rise, nine in 10 executives in product design, marketing, and sales believe agents will manage at least one process in their functions within the next five years.
Still, trust remains a barrier. About 71% of companies said they cannot yet fully rely on autonomous agents, and fewer than half have robust AI governance policies in place.
