💡 In CEO rooms lately, there’s a quiet shift underway when it comes to AI. Leaders aren’t talking about AI strategy anymore — they’re talking about agentic operating models. Not the next app to pilot, but the next way their organizations will actually work. The learning curve is steep, but the dialogue sounds different: ❓“How do we make sure people trust the output?” ❓“Who’s accountable when an agent makes a call that used to be a manager’s?” ❓“Where’s the 10x productivity without losing judgment?” The latest McKinsey research from my QuantumBlack, AI by McKinsey colleagues captures this moment perfectly. One story stood out that stood out to me. A bank reimagined how its relationship managers create credit-risk memos. What once took weeks of data gathering now happens in hours: agents pull data, draft sections, and flag low-confidence areas for human review. Productivity rose 20–60%, decisions improved, and people focused on what truly required judgment. (More on that example and full playbook in: Seizing the Agentic AI Advantage) AI isn’t replacing leadership — it’s redefining it. For CEOs, that means: 1️⃣ Start with a lighthouse workflow, not a single task. 2️⃣ Design for trust and accountability. 3️⃣ Appoint an “agent leader” to make human + machine collaboration a discipline. 4️⃣ Be explicit about the judgment line — what you’ll never delegate. This is the hard, messy, exciting work of leadership in the agentic age. Kudos to my colleagues Alex Singla, Aamer Baig, Oana Cheta, Alexander Sukharevsky, Lari Hamalainen, Olli Salo, Pallav Jain, Raghav Raghunathan, and Sandra Durth for leading the way. (Link to the main piece below: The Change Agent: Goals, Decisions, and Implications for CEOs in the Agentic Age)
Nice list Carolyn Dewar. I would add another…what agentic workflows are critical to “own”? By definition, these workflows require orchestration of models, sub-agents and bots with disintermediation and IP leakage potential at various stages. Finding the right blend of internal and 3rd party solutions is an important concern.
This nails it. We’ve officially entered the era where AI isn’t just automating tasks. It’s reshaping the architecture of work itself. The smartest CEOs I’m seeing aren’t asking “what can AI do?” anymore; they’re asking “how do we lead differently because of it?” The real differentiator now is trust, accountability, and human judgment. That’s the frontier where leadership is being redefined. Messy, yes, but wildly generative.
Carolyn Dewar Ceo has to now become the throughput of Trust where collaboration of human wisdom and agentic Ai happens...
Fantastic paper, and a great example of how differently CEOs and CHROs need to think about talent acquisition and priorities for the C-Suite and Board.
Insightful post Carolyn Dewar Perfectly captures how the AI conversation is shifting from strategy to agentic operating models. It’s no longer about tools, but how organizations work, build trust and define accountability. AI isn’t replacing leadership; it’s redefining it.
Brilliant.
Productivity is increasing definitely . But at what cost?. Human are becoming food eating machines with every passing day. Views are my very own,
AI is an amazing tool. It's certainly not perfect and human oversight in my opinion is key to verify the output. If you can get that right the improvements to productivity are enormous.
McKinsey Sr Partner | NYT-Bestselling Author | CEO Practice leader | Strategy, Growth, Organization Effectiveness & Transformation
2wLink to article here: https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-change-agent-goals-decisions-and-implications-for-ceos-in-the-agentic-age