AI Agents in Action Foundations for Evaluation and Governance 2025

Foreword In recent years, organizations have moved beyond predictive models and chat interfaces to experiment with artificial intelligence (AI) in more transformative ways. AI agents are now emerging as integrated collaborators in business, public services and everyday life. The adoption of AI agents could bring significant gains in efficiency, altered kinds of human-machine interaction and the advent of novel digital ecosystems. This transition faces multiple obstacles that need to be addressed. Moving from models to agents represents more than a technical milestone and requires organizations to rethink how they design, evaluate and govern advanced agentic systems. Many companies are now questioning what agents can accomplish alongside the practical steps needed to adopt and deploy them safely, responsibly and effectively. This paper was developed to help answer those questions. By mapping the evolving foundations of agentic systems, classifying their roles, identifying new ways to evaluate them and outlining progressive governance approaches, the paper offers practical guidance for leaders navigating adoption in real-world contexts.Through the AI Governance Alliance, the World Economic Forum and Capgemini are advancing this subject in collaboration with the AI community, signalling that now is the time to prepare for an agentic future. If adopters start small, iterate carefully and apply proportionate safeguards, agents can be deployed in ways that amplify human capabilities, unlock productivity and establish a foundation for more complex multi-agent ecosystems to emerge over time. Unless a careful and deliberate approach to adoption is adopted, untested use cases could outpace oversight and lead to misaligned incentives, emergent risks and loss of public trust. As with any transformative technology, the opportunities presented by AI agents must be accompanied by a responsibility to guide their development and deployment with care. Through cross-functional efforts and collaborative governance, AI agents can be integrated in ways that amplify human ingenuity, promote innovation and improve overall quality of life. This paper is a step in that direction, offering guidance to help early adopters navigate the complex and often uneven path of AI agent adoption. Roshan Gya Chief Executive Officer, Capgemini InventCathy Li Head, Centre for AI Excellence, Member of the Executive Committee, World Economic Forum AI Agents in Action: Foundations for Evaluation and GovernanceNovember 2025 AI Agents in Action: Foundations for Evaluation and Governance 3