Seizing the agentic AI advantage

The report "Seizing the Agentic AI Advantage" by McKinsey & Company (June 2025) examines the generative AI paradox: although around 78% of companies have adopted this technology, more than 80% see no measurable benefits and only 1% consider their AI strategy mature. This highlights the gap between widespread adoption and real business impact.

The study shows that many firms focus on horizontal use cases such as copilots or chatbots, which provide diffuse and hard-to-measure improvements. By contrast, vertical use cases designed for critical business functions hold greater transformative potential, but 90% remain stuck in pilot stages due to technical, organizational, and cultural barriers.

The proposed solution is AI agents: autonomous systems with memory, reasoning, planning, and execution that act as proactive collaborators rather than reactive tools. Unlike traditional generative AI, they can automate complex workflows involving multiple steps and systems, improving efficiency, generating new revenue streams, and boosting agility. Reported examples show cost and time reductions of up to 50% and significant gains in customer service and data analysis.

To capture these benefits, companies must redesign processes from scratch instead of simply inserting agents into existing workflows. Their strengths — parallel execution, real-time adaptability, and large-scale personalization — require a rethinking of tasks. To manage them at scale, McKinsey introduces the concept of an agentic AI mesh, a modular architecture combining customized and commercial agents to ensure interoperability and security.

Adopting agents also raises challenges: controlling autonomy, preventing redundancy, and establishing robust governance. The report proposes four pillars for success: people (new skills and roles), governance (aligning autonomy with business goals), technology (scalable infrastructure), and data (quality and productization).

Aimed at CEOs and digital transformation leaders, the report concludes that the era of experimentation is over. The future of AI in business lies in adopting agents as active collaborators, redesigning operating models, and unlocking sustainable value.

Key Points

• The report describes the generative AI paradox: broad adoption but little real business impact.
• Over 78% of firms use generative AI, but only 1% have a mature strategy with clear benefits.
• Most investment goes to horizontal copilots and chatbots with limited, diffuse gains.
• Vertical use cases in critical functions hold more potential but 90% stall at pilot stage.
• AI agents are the proposed solution: autonomous systems with memory, reasoning, and execution.
• Agents can automate complex processes, cut costs by up to 50%, and enhance customer service.
• To succeed, companies must redesign processes from scratch, not just add agents.
• McKinsey introduces the agentic AI mesh, a modular architecture for large-scale integration.
• Adoption requires four pillars: people, governance, technology, and data.
• The report concludes: experimentation is over; firms must act now to achieve sustainable value with AI agents.