What Must CEOs do to Lead in the Age of Agentic AI?

Generative AI tools exploded into the mainstream earlier in the 2020s, but the spotlight is now shifting toward AI agents – autonomous systems that can plan, learn and act to achieve business outcomes. 

While headlines focus on novelty, the real story is unfolding inside companies: how these agents are beginning to reshape workflows, decision-making and even entire operating models.

A report by McKinsey & Company highlights how this brings both urgency and opportunity for CEOs.

It says that while early adopters are seeing accelerated project delivery, lower operating costs and scalable productivity, achieving enterprise-level impact requires strategic leadership, not just experimentation.

According to McKinsey, CEOs can lead the transformation from workflows to agent-first organisations ready for the future – but how?

Implementing AI agents

AI agents are gaining momentum – and McKinsey & Company say they are delivering real business value.

Its research says that initial implementations show AI agents can accelerate delivery timelines by 40–50%, cut costs by over 40% and improve quality – especially when deployed at scale. 

In one case, a bank tasked just five humans to supervise 100 AI agents and halved the time and labour required for tech modernisation.

But not all agents are the same. To unlock their full potential, McKinsey says CEOs must understand the difference between these two agents:

1. Agentic Labour – these are tools that assist or automate existing tasks and include:

• Individual augmentation: Agents help people do their work faster – e.g., drafting documents or writing code – boosting personal productivity by 20-30%.
   
• Workflow automation: Agents automate parts of structured processes (like financial reporting or customer service), cutting cycle times by up to 40%.

2. Agentic Engine – these are systems redesigned around teams of agents:

• Functional agentic workflows: Entire functions, like supply chain or finance, are rearchitected to run on agent teams – removing bottlenecks and improving outcomes.
   
• Cross-functional systems: Complex, end-to-end workflows (e.g., customer journeys or financial cycles) are managed autonomously, delivering up to 80% cost reductions and faster time to market.

The report says that for CEOs, the shift from isolated tools to agent-first systems requires more than just tech – it demands operating model transformation and strategic leadership.

The CEO agenda

As AI agents rise in popularity, McKinsey outlines a bold two-to-three-year roadmap for CEOs looking to lead their organisations through agentic transformation. 

First year

It says that the early phases should include building understanding, creating momentum and laying the foundations so AI agents can work at scale.

The focus should be on driving down operating costs of existing activities – considering a goal of around 10% – and encouraging enterprise-wide adoption.

By the end of year one, CEOs should expect at least 25–50% of employees to be working with AI agents daily.

Second year

Year two marks the shift from experimentation to measurable impact, where automation scales across value streams and customer journeys become reimagined. 

McKinsey gives the example of AI-driven order-to-cash systems – the goal could be to automate more than 70% of transactions across all channels.

To achieve this, CEOs must drive transformation at scale – establishing a central “agentic factory” to govern design standards, performance KPIs and control sprawl. 

Platforms for managing agents, data infrastructure, and agent trust will be critical. Crucially, this evolution demands a shift in talent strategy. 

Every employee must move from using AI tools to managing them, with new roles like “agent orchestrator” and KPIs tied to agent collaboration.

Third year 

By year three, automation becomes business-as-usual: over 90% of key workflows are agent-led, professionals command multiple agents, and FTE needs fall by up to 40% in some domains. 

“Agentic automation is driven in 90% of key value streams,” McKinsey notes at this stage of the roadmap.

Long term, CEOs must reimagine the business model entirely. McKinsey says they should ask themselves key questions like: 

• How will agentic AI affect my business model, challenge our existing sources of differentiation and create new ones?
• What should my transformation and investment roadmap look like to both meet near-term business goals and establish the right foundations to develop and scale transformational change?
  

In this new era, the pace of change in the realm of artificial intelligence is dizzying. McKinsey says decisive and thoughtful action is the way leaders open new opportunities for productivity and growth.