Leadership in the Age of AI: The Five Skills Needed to Lead People and Agents

As we continue to work and live in a constantly disruptive world, leadership is more important than ever.

In Fortune’s 2025 Return to Leadership (ROL100) Report, companies in the top 25 had a median EBITDA per employee of $180,000, while those in the bottom 25 ranking had only $44,000.¹ This shows that leadership drives higher performance.

As AI reshapes the future of work, leadership must evolve from managing execution to guiding people and the business. Five distinct leadership capabilities are essential in the age of AI. These are critical to truly lead a human AI hybrid workforce that moves beyond the use of generative AI into the agentic era. Neglecting the development of these skills can determine whether AI becomes a powerful asset to your workforce or fades as a short-lived trend that fails to justify the investment. In such cases, employee cynicism may linger for months or years, setting an enterprise back significantly on its strategic journey.

The Leadership Shift: From Execution to Elevation

AI is no longer an emerging technology—it is a true business imperative. As AI systems and agents increasingly take on operational tasks, they are changing the demands placed on humans. In a world where decisions are accelerated and information is seemingly infinite, leadership must shift from directing work to enabling people and organizations to thrive in speed and complexity.

In response, organizations are reevaluating what effective leadership looks like. We conducted a benchmarking study of 15 global organizations across seven sectors, including AI-first firms, and applied Generative AI to review AI’s impact on leadership through academic and thought leadership studies. Our aim was to uncover the common themes shaping high-performing, future-ready leadership models.

A clear pattern emerged. While companies may differ in language or structure, the capabilities they prize in leaders in the age of AI converge around five essential skills.

1. Coaching: Building Capability Over Control

AI handles tasks. People drive performance. As the nature of work changes, leaders must focus less on oversight and more on building capabilities. When humans delegate to other humans and outcomes fall short, we often blame the person. With agents, failures may stem from integration gaps, tool/model selection or contextual grounding, but accountability still falls on humans—either the task was not structured properly, or the AI was not set up to succeed (e.g., due to data quality issues, vague requests or ill-defined business processes).

In our benchmarking, 75% of organizations identified coaching or team empowerment as a critical leadership behavior. This represents a shift from directing tasks to fostering growth, enabling people to solve problems, collaborate effectively and step into leadership roles themselves.

What effective coaching for humans looks like:

• Delivering timely, constructive feedback
• Encouraging autonomy through stretch assignments
• Creating psychological safety for experimentation and learning

What effective coaching for agents looks like:

• Defining tasks and requests with exceptional clarity
• Creating an iterative improvement cycle with continuous feedback between humans and agents
• Examining human input for phrasing, data availability and potential biases when responding to hallucinations

These practices reduce rework time, improve task accuracy and raise adoption and satisfaction scores when humans delegate to agents.

2. Creativity: Asking Better Questions, Not Having All the Answers

AI excels at recognizing and replicating patterns. But it’s human curiosity and creativity that define the problems worth solving—and the possibilities worth pursuing. In top-performing organizations, creativity is not limited to innovation departments; it is expected of leaders across functions, creating environments where anyone can share ideas and make an impact.

What effective creativity for humans looks like:

• Reframing challenges to find new solutions
• Embracing fast, small-scale experimentation
• Rewarding bold thinking

Organizations only win when humans focus on the right things and processes are designed to enable those focus areas. Otherwise, AI simply accelerates existing bottlenecks. If a process is broken or siloed, AI will not remove the inefficiency—it will amplify it.

What effective creativity for agents looks like:

• Redesigning workflows so AI augments collaboration instead of replicating human friction
• Addressing human-controlled factors when AI falls short (e.g., data quality, process structure, backend systems)
• Exploring additional opportunities where AI can solve problems with existing or attainable input

3. Agility: Leading Through Change at Speed and Scale

The pace of technological change is exponential. Agility – the ability to pivot quickly while maintaining long-term focus – is a necessity. Leaders must not fear mistakes; instead, they should surface and correct them before they become critical.

Organizations we studied are embedding agility into leadership development by shortening decision cycles, decentralizing authority and encouraging iterative strategy-setting.

What effective agility for humans looks like:

• Making decisions at pace amid ambiguity
• Re-prioritizing resources rapidly
• Keeping teams aligned as opportunities arise and priorities shift

What effective agility for agents looks like:

• Institutionalizing monitoring and scoring practices through a model validation framework that ensures quality doesn’t drift as models or data evolve
• Establishing systems to detect, set thresholds and escalate hallucinations, ensuring errors trigger updates and pattern recognition
• Identifying breakdowns in real time and addressing infrastructure gaps

4. Connectivity: Reinforcing Trust, Inclusion and a Shared Culture

Automation and hybrid work risk isolating individuals and fracturing culture. Effective leaders counteract this by being deliberate connectors—bridging silos, reinforcing values and fostering collaboration. Poorly managed hybrid work can erode psychological safety, trust and wellbeing.

Connectivity between both humans and agents is also critical. This includes reinforcing responsible AI practices and governance, embedding trust and compliance into human-agent collaboration.

What effective connectivity for humans looks like:

• Building communities for people to discuss key issues and change adoption challenges
• Linking expected behaviors (i.e., values) to business and customer outcomes
• Breaking down silos by actively looking past/finding ways around organizational boundaries

What effective connectivity for agents looks like:

• Embedding guardrails (e.g., responsible AI frameworks, explainability standards)
• Creating new ways for humans and agents to collaborate as technology evolves
• Clarifying how AI outputs are validated and how accountability is maintained

5. Sense-Making: Providing Clarity in Complexity

As AI accelerates change, ambiguity increases. Leaders must now act as translators—helping people understand shifting priorities, evolving workflows and emerging opportunities. They must also create clarity that agents require a period of trial, adjustment and refinement before performing optimally. Demonstrating confidence and patience with agents is essential for adoption.

High performing organizations emphasize sense-making as a core skill: the ability to communicate complex ideas simply, instill confidence and create meaning amid uncertainty.

What effective sense-making for humans looks like:

• Framing AI and transformation as opportunities
• Sharing direction with conviction and honesty
• Creating emotional clarity during disruption

What effective sense-making for agents looks like:

• Providing agents with clear context and allowing for iterative improvement cycles—just as with onboarding a new team member
• Being realistic about AI’s capabilities and timelines
• Remembering that humans and agents serve a common purpose

Growing the Five Leadership Capabilities

All companies invest in leadership development, but few achieve real behavior change. New skills require new methods. Too often, companies rely on day-long sessions focused on one capability, which are often academic and not directly applicable. This is not a failure of intent—it’s a failure of design.

Several of our clients have moved away from such sessions. They’ve learned that to grow leaders, program design must be:

• Framed in the leader’s flow of work (i.e., relevant to their role and company)
• Experiential, allowing practice in a safe setting (e.g., using AI)
• Practical, with takeaways leaders can apply immediately

Example: “Hot Seat” Exercise

To build sense-making, one effective method places leaders in the “Hot Seat” to deliver an inspirational, clear and effective message.

• Leaders select a real-world prompt (e.g., announcing a change, inspiring a team)
• They prepare for 2-3 minutes and deliver a three-minute talk to a small peer group
• Each peer provides feedback on clarity, tone and engagement
• The leader applies the feedback and delivers an improved 90-second version
• Peers give additional feedback
• The leader reflects on lessons learned and how to apply them with their team

This exercise builds confidence, presence and real-time clarity—skills that passive training cannot teach.

The Impact on Organizations and Outcomes

The AI era brings new demands from employees. They expect increased communication, clarity on evolving roles, guidance on needed capabilities and help in building internal connections. If these demands go unmet, AI programs will fail.

According to MIT’s latest research, only 5% of generative AI pilot programs drive rapid revenue acceleration—meaning 95% fall short.² While most pilots fail without strong leadership, those that succeed demonstrate outsized impact, from measurable productivity gains to new client offerings. For example, consulting benchmarks show firms that redesigned workflows with AI saw 2.5x higher revenue growth and 2.4x productivity gains compared to peers.³

These successes show that developing the right leadership capabilities is essential in the age of AI. For CEOs and boards, success isn’t about tools, but about leaders who can embed judgment, validation and culture into every workflow—ensuring AI amplifies, rather than erodes, organizational trust and performance.

The five capabilities are imperative—not just for managing humans, but for thriving in a hybrid world of humans and agents.

¹ Purpose-driven leadership drives growth—and these Fortune 500 titans prove it | Fortune

² The GenAI Divide: State of AI in Business 2025. MIT Decentralized AI

³ Reinventing Enterprise Operations with Gen AI. 2024. Accenture