Share with your CHRO
McKinsey’s case for HR’s dual mandate in the AI era puts a sharp number on the credibility gap: 88 percent of companies now run AI in at least one business function, but only 39 percent report any material EBIT contribution. The argument is that HR must simultaneously redesign work architecture across the enterprise and transform its own function first, because an HR team still running manual onboarding has no standing to demand that sales or finance reimagine their workflows. McKinsey estimates AI could generate $150 billion to $200 billion in annual global value in HR alone, with two-thirds of HR processes partially or fully automatable.
What this means for your business
The 88-to-39 gap is the real story here. Most organizations have deployed AI broadly and captured almost nothing, and the failure pattern isn’t a technology shortfall, it’s a workflow and capability shortfall. CHROs sitting inside that gap face a specific positioning choice: HR can remain an administrative function that advises on change while others experience it, or it can become the proof-of-concept layer the rest of the enterprise watches. Which side of that line your function sits on determines whether you’re consulted or ignored when operating model decisions get made.
The “messy middle” figure McKinsey cites deserves more attention than it usually gets. Seventy-six percent of jobs are neither fully automatable nor untouched by agentic AI, meaning the dominant workforce challenge isn’t replacement, it’s role redesign at scale. That’s harder than a reskilling program and harder than a layoff plan, because it requires simultaneously modeling task distribution across humans and AI agents, rewriting job architectures, and building governance for decisions that now involve non-human actors. The organizations that have tried to run this as a technology project, leaving HR on the side, are the ones producing the 39 percent.
The piece is authored by McKinsey consultants who sell the exact transformation programs they’re prescribing, which tilts the timeline toward urgency and the scope toward enterprise-wide engagements rather than targeted interventions. That incentive is worth naming, but it doesn’t invalidate the core diagnosis. Where I’d push back is on the “lighthouse” framing as a precondition: some HR functions will move faster by embedding in a business unit’s AI deployment and learning in context, rather than self-transforming in isolation first. The falsification condition is simple enough. If HR functions that go last on internal transformation still consistently outperform on enterprise AI ROI, the dual mandate thesis collapses.
Based on reporting from AI in HR transformation: A dual mandate, originally published 2026-06-08 03:00:00.
