Generative AI isn’t just an experiment anymore—it’s moving into the fabric of enterprise operations at warp speed. The latest Capgemini Research Institute report, Harnessing the value of AI: Unlocking scalable advantage, shows that organizations are accelerating adoption and rewriting team structures to accommodate AI as a genuine “colleague.”
According to the survey of more than 1,100 executives, nearly 6 in 10 organizations expect AI to act as an active team member or supervisor for other AI systems within the next 12 months—a big jump from 44% today. That shift isn’t just technical; it’s organizational. Two-thirds of companies admit they’ll need to restructure teams to enable smoother human-AI collaboration.
Scaling Up Fast—But at a Cost
The numbers tell a story of runaway momentum. Thirty percent of organizations are now fully or partially scaling Gen AI—a fivefold increase since 2023. Nearly every enterprise surveyed (93%) is at least experimenting with the technology in 2025. Telecom, consumer products, and aerospace/defense are leading adoption, especially in customer operations, IT, marketing, and risk management.
But speed comes with turbulence. While 88% of companies boosted Gen AI investment by an average of 9% in the past year, many report a nasty surprise: cloud “bill shock.” As AI workloads balloon, over half of organizations say consumption costs are outpacing expectations. To blunt the impact, some are pivoting to small language models (SLMs), which promise lighter compute requirements and more cost efficiency than massive LLMs.
Rise of AI Agents and Multi-Agent Systems
Capgemini’s data also highlights a boom in AI agents—autonomous bots trained to execute specific business tasks. Nine in ten executives in product design, marketing, and sales say they expect AI agents to manage at least one process in their function within five years.
The next frontier? Multi-agent systems, where interconnected agents collaborate. Nearly half of enterprises already piloting agents are experimenting with multi-agent setups. And 38% believe AI agents will evolve into self-learning systems with minimal human oversight in the next 3–5 years.
Still, there’s a trust gap: 71% of enterprises say they can’t fully trust autonomous AI agents yet. Governance is patchy too—fewer than half have formal AI governance policies, and many of those aren’t consistently enforced.
Expert Take: Tech Isn’t Enough
Franck Greverie, Capgemini’s Chief Technology & Portfolio Officer, warns that scaling AI isn’t just about plugging in models: “Rapid adoption doesn’t necessarily translate into tangible ROI. Enterprises need a trusted, compliant data foundation—and a new operating model with balanced human-AI chemistry.”
The Bottom Line
Gen AI is scaling faster than most tech waves before it. But companies face a paradox: adoption is accelerating while governance, trust, and cost controls lag behind. Enterprises that crack the human-AI operating model—balancing speed with oversight—may be the ones to turn experimentation into competitive advantage.
The full report, based on a survey of 1,100 executives across 15 countries, is available from the Capgemini Research Institute.
Join thousands of HR leaders who rely on HRTechEdge for the latest in workforce technology, AI-driven HR solutions, and strategic insights
