McKinsey's Bob Sternfels: Half of AI's Value Comes from Organizational Change, Not Technology

McKinsey's Global Managing Partner Bob Sternfels appeared on Harvard Business Review's podcast and said the quiet part out loud: for large enterprises, half or more of AI's value comes from flattening the organization, not from the technology implementation itself. The insight reframes the entire enterprise AI conversation from a technology problem to an organizational design problem.

The Core Insight

Sternfels' central argument is straightforward: AI only works when you break down silos. The technology is necessary but insufficient. The real unlock comes from rethinking how the organization is structured around the workflows that AI is being applied to.

The Mortgage Example

Consider a mortgage process. It has multiple steps: origination, credit scoring, collection, and after-service. In a traditional bank, each of those steps is a separate department. That departmental structure exists because humans needed specialization and handoff boundaries.

But if AI can handle the end-to-end workflow, the question becomes: why do you have four or five departments in a single process? Can you break those walls down and let AI operate across what used to be organizational boundaries?

What This Means in Practice

Sternfels frames the challenge in three layers:

Layer	Question	Difficulty
Strategy	What is the AI strategy?	Medium
Implementation	How do you implement the technology?	Medium
Organizational Redesign	How do you rewire the organization to realize the value?	Hardest

The third layer -- organizational redesign -- is where most enterprises are struggling. When you get it right, the CFO and CIO end up on the same page. But McKinsey is finding that this alignment is harder and takes longer than people expected.

The Bigger Picture

Sternfels believes we are entering a period where enterprises will fundamentally change themselves. The potential is enormous, but getting there requires patience. The companies that treat AI as a technology implementation project will capture a fraction of the value. The ones that use AI as a catalyst for organizational transformation will capture the rest.

This tracks with McKinsey's broader positioning: the firm now describes its workforce as "40,000 humans and 20,000 agents," signaling that even the world's largest consultancy is restructuring itself around AI, not just advising clients to do so.

Key Takeaways

• Half or more of enterprise AI value comes from organizational change, not technology
• AI works best when it breaks down departmental silos in end-to-end workflows
• Organizational redesign is the hardest layer and is taking longer than expected
• When CFO and CIO alignment happens through restructuring, the value materializes
• Enterprises that treat AI as a technology project will underperform those that use it as an organizational catalyst

Resources

• Where McKinsey -- and Consulting -- Go From Here -- HBR IdeaCast podcast episode with Bob Sternfels
• McKinsey on Agentic Organizations -- McKinsey's framework for AI-native org design