AI agents offer major value but trust and data gaps remain
Trust in agentic AI drops as firms move from pilots to deployment.
AI agents could drive up to $450 billion in economic value by 2028, according to new research by Capgemini. The gains would stem from increased revenue and cost savings, but not all businesses are equally positioned to benefit.
The report found that companies with scaled AI agent deployment could realise $382 million in gains over three years. By contrast, lower-scale adopters are forecast to see just $76 million in the same period.
Despite the potential, only 2% of organisations worldwide have fully scaled AI agents. In the UK, the figure drops to just 1%, highlighting a significant adoption gap.
Nearly half of UK firms are still piloting AI agents or exploring use cases. Only 19% have partially implemented them, while most respondents (93%) agree scaling AI agents would offer a competitive edge within the next year.
Customer service, IT and sales are expected to see the biggest benefits from AI agent deployment. Yet as more companies shift from pilots to implementation, trust in agentic AI has declined sharply.
Only 27% of global respondents now express trust in AI agents, compared with 43% last year. In the UK, trust fell from 50% to 32%, though the level remains slightly above the global average.
Franck Greverie, Capgemini’s Chief Technology Officer, said trust must be built into AI from the outset. Responsible development practices, grounded in ethics and safety, are essential for long-term adoption.
A lack of AI infrastructure remains the top barrier to scaling. Four in five organisations say their infrastructure is immature, and fewer than one in five report high data readiness.
AI’s performance remains tied to data quality. Internal AI literacy is also low across sectors, while privacy concerns add further friction to scaling efforts.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!
