The Coordination Layer Is Collapsing. CHROs Are Last to Know.
The most important AI story this week isn't about a product launch or a funding round. It's about what happens when AI quietly absorbs the connective tissue of how organizations actually run.
THE DEEP CUT
Here's a question nobody in HR is asking loudly enough: if AI handles coordination, what exactly is a first-line manager for?
SHRM's AI+HI Project is reporting what practitioners in Big Tech have been watching for months: organizations are flattening, spans of control are expanding, and the stated reason is that AI is absorbing the coordination tasks that justified management layers in the first place. Status updates. Meeting routing. Escalation triage. Information packaging. The administrative scaffolding that made a manager necessary for every six to eight people.
Block just demonstrated the scale of this math. CEO Jack Dorsey eliminated 40% of the company's workforce and explicitly framed it as a proactive restructuring around AI-embedded teams. That 40% figure is now circulating in board decks everywhere. CFOs are doing the math on their own organizations. The question forming in audit committees right now is: we've been deploying AI tools for 18 months, where is the productivity dividend in headcount? If you don't have a workforce planning model that connects AI adoption to capacity, you will answer that question reactively, in someone else's framing, with data you pulled the night before the meeting.
Here's the part that isn't getting enough air: every org restructuring I've seen over twelve years gets the denominator right and the numerator wrong. Finance targets a span ratio, maybe one manager per twelve instead of one per seven. Headcount comes out. The org chart gets flatter. And then eighteen months later, leadership is puzzling over collapsed engagement scores, missed execution on strategy, and a cohort of middle managers who are technically still there but clearly demoralized and directionless. The coordination work got cut. The new work never got defined.
This is the design failure hiding inside the "Great Flattening" narrative. When AI absorbs coordination, it creates a vacuum where management bandwidth used to live. The organizations getting this right are asking what fills that space. Strategic coaching, for one. But also the messier stuff: resolving ambiguity when priorities conflict, holding people accountable to culture when nobody's watching. The judgment calls that require knowing the team, knowing the context, and sitting with two conflicting priorities at once. That's what AI cannot do, and that's exactly what gets crowded out when managers are buried in status updates.
The organizations getting this wrong are treating flatter as a cost outcome rather than a design outcome. They're removing layers, handing each surviving manager fifteen direct reports, and calling it transformation. It isn't. It's cost-cutting with a narrative wrapper.
The concrete action here is structural, not analytical. CHROs need to audit current span-of-control ratios by function, identify where AI tooling is already absorbing coordination work, and then rewrite the first-line manager job architecture before Finance writes it for you with a single number. That rewrite isn't a job description exercise. It's a value proposition exercise: what does a manager in this function exist to do in a world where coordination is automated? If you can't answer that cleanly, you don't have a manager design. You have a legacy reporting structure with a new budget pressure applied to it.
Fair counterargument: not every function is ready for expanded spans. The research on span limits in high-complexity roles is real. Managing fifteen engineers doing novel technical work is categorically different from managing fifteen customer service representatives with AI-assisted workflows. Span-of-control optimization is not one-size-fits-all.
That's exactly why the CHRO needs to own the design, function by function, before a blunt financial target lands from above.
QUICK SIGNALS
192,000 HR Jobs Face High Automation Risk with No Technical Barriers
SHRM research identifies 192,000 U.S. HR positions that are already at least 50% automated with no regulatory or client-preference barriers preventing further displacement. The key phrase: "no nontechnical barriers." The obstacle to automation in these roles isn't legal protection or relationship dependency. It's organizational will. CHROs who are advising other functions on AI displacement while leaving their own team unexamined have a credibility problem they may not see coming.
Monster.com Founder Launches AI-Native Talent Marketplace Boomband
Jeff Taylor, founder of Monster.com, launched Boomband, an AI-native platform that replaces resumes and job descriptions with AI-generated dossiers, backed by a $4M seed round. Taylor's diagnosis is credible: current AI in recruiting is mostly noise layered on broken process. The harder problem is that replacing a 70-year standard requires candidates and employers to change simultaneously. That two-sided coordination problem is where most TA tech stalls, regardless of how good the product is. Watch the traction, not the launch.
THE HUMAN ANGLE
The question I keep turning over this week: we talk about AI eliminating jobs, but the more interesting disruption is AI eliminating reasons that jobs exist.
A middle manager's role didn't start as a job title. It started as a solution to a coordination problem. Someone needed to route information, triage decisions, and translate strategy into daily work for a team. The job crystallized around those tasks. Now AI is dissolving the tasks, and organizations are treating the job title as if it's what needs to go, rather than the underlying problem that needed solving.
The same thing is happening in HR operations, in recruiting, in L&D. The transactional tasks that justified whole role categories are getting absorbed quietly. What's left are judgment calls. Someone has to navigate ambiguity. Someone has to hold relationships intact when priorities shift. Those demands are harder to staff for, harder to train for, and harder to measure. Which is exactly why most organizations won't redesign for them. They'll just cut the denominator and call it progress.
The CHROs who come out of the next 18 months with intact functions and organizational trust will be the ones who got ahead of this question: what does each role on my team exist to solve? Not what do they do, but what problem would go unsolved without them? That's a hard question to ask about your own team. It's also the only question that matters right now.
-- Alex