How Meta and Block Are Splitting Up Management Work

Meta and Block Are Reshaping What Managers Do

Meta and Block are restructuring their organizations on the assumption that AI will be part of the workflow. Alongside efforts to reduce the number of managers, I’m seeing both companies split up traditional management duties and hand them off to other roles.

According to Business Insider, Meta has launched a pilot program within Reality Labs — a team of roughly 1,000 people — that’s overhauling job titles and roles. At Block, Jack Dorsey said he wants to cut the management layers between himself and his roughly 6,000 employees down to just 2-3 this year. He’s even floated, as an ideal endpoint, a structure where every employee reports directly to him. It’s worth distinguishing between what’s already been implemented and what remains a stated goal.

What caught my attention is how the two companies are dividing up a manager’s job. Collecting and relaying status updates, setting priorities, and evaluating and developing people are three genuinely different functions. Even if you cut out layers of reporting, someone still has to own each of these tasks.

Reporting, evaluation, and talent development go to separate roles

Meta’s team is introducing three new titles: “AI Builder,” “AI Pod Lead,” and “AI Org Lead.” Not everyone becomes an AI Builder. The Pod Lead handles day-to-day work progress, while the Org Lead manages performance reviews and promotions with AI assistance. The reported internal document notes that overall team headcount won’t change under this reorganization.

Block envisions a structure where AI takes on more of the work of relaying information and coordinating tasks. In a March 31 essay, Jack Dorsey and Sequoia’s Roelof Botha argued that if AI continuously tracks a company’s code, design work, discussions, and progress records, it can reduce the number of steps needed to pass reports up through managers. The AI system described in the essay also includes an “intelligence layer” that assembles financial features to match customer needs.¹

The plan is to split human roles into three types: subject-matter experts, DRIs² responsible for specific tasks, and “player-coaches” who combine hands-on work with mentoring colleagues. This reminds me of “kkochi coaching” — a peer-mentoring program I named after kkochi, Korean-style skewers, that I’ve been running since 2023. Even as AI use expands, someone still needs to help people grow.

This plan has to be read alongside a massive workforce reduction. In its February 2026 shareholder letter, Block announced it would cut headcount from over 10,000 to under 6,000. More than 4,000 employees are being asked to leave or are in related discussions. This comes even as 2025 gross profit rose 17% year over year. Dorsey has argued that smaller teams can accomplish more using AI tools, but the scale of the cuts alone doesn’t prove that claim.

Even after the titles change, if the same people keep making the same decisions and receiving the same reports, actual operations may not shift much at all. What matters is comparing who decides what — and who is accountable for it — before and after the reorganization.

We Should Separate “Fewer Managers” Data From AI’s Actual Effect

In Indeed data cited by Business Insider, middle-manager job postings in 2025 fell 12.3% from the previous year. But overall job postings also declined, so this figure alone can’t isolate AI’s impact. In Korn Ferry’s 2025 survey, 41% of respondents said management layers had shrunk within their own organizations. That’s a survey asking employees about organizational changes they’ve experienced.

In 2024, Gartner projected that by 2026, 20% of organizations would use AI to flatten hierarchies, eliminating more than half of those organizations’ middle-management positions. This projection shouldn’t be read as a statistic of reductions that have already happened.

Cost-cutting and faster decision-making are also reasons for trimming management layers. What AI actually did in that process needs to be verified separately. We need to look at headcount reductions, the workload on remaining employees, and organizational performance all together.

Zappos Also Had to Redefine Authority and Responsibility

Zappos’s rollout of Holacracy³ in the 2010s is worth comparing here too. It was an attempt to distribute authority by role and team instead of relying on traditional manager titles.

The change, pushed by then-CEO Tony Hsieh, was meant to give employees more autonomy. But people had to learn to understand roles and decision rights all over again, and some employees left during the transition. Simply eliminating manager titles didn’t automatically solve the problems of how to collaborate and who’s accountable for what.

Even in an organization built around autonomous work, you still need someone to approve budgets, someone to make the call when conflicts arise, and criteria for evaluating results. If that authority moved from managers to roles or teams, members need to be able to see exactly how far it moved.

Harvard Business School professor Linda Hill explained in a Business Insider interview that small teams with diverse expertise can drive innovation, but you still need a human leader connecting collaboration across teams. I read this as saying that autonomy within a team and coordination between teams have to be designed together.

Block expects that having AI handle the work of gathering and coordinating information can ease these old difficulties. Its essay also acknowledges that the transition is still in its early stages and that some of it will involve trial and error. How well that system actually performs this coordination work, and how much of it, remains to be seen.

Oswarld’s Lens

I see this reorganization as both an experiment in how work gets done and a way of explaining the company’s future to investors. I can’t say exactly how the two motives are mixed together.

Studying corporate cases in my MBA program, I came to think that organizational restructuring often serves two purposes at once: actually changing how operations run, and signaling direction to the market. I’m reading Block’s explanation through that same lens. Improved profitability from lower labor costs and improved speed and quality of work thanks to AI need to be evaluated separately.

AI tools can help summarize reports and track progress. But allocating resources or adjusting priorities requires judgment that goes beyond organizing information. And how much responsibility can be handed off for resolving conflicts between colleagues, understanding the context of work, or supporting career development — that needs to be tested role by role.

Patty McCord, Netflix’s former chief talent officer, made a similar point in the same interview: she expects AI to replace certain tasks rather than entire people. Block’s player-coach concept reflects that same distinction. It reduces the burden of relaying information while preserving a role that involves hands-on participation in the work and mentoring colleagues.

What I most want to see clarified is decision-making authority and accountability. Even when work is divided across multiple roles, it’s possible to designate who has final say. Conversely, having a title means little if authority and responsibility remain unclear. I think the company needs to spell out who approves the priorities AI suggests, and who corrects it when its judgment goes wrong.

Closing

I’d measure the success of this reorganization on three fronts.

• Whether decisions actually got faster now that there’s less reporting and fewer meetings
• Whether the remaining staff can handle both hands-on work and mentoring peers at the same time
• Whether authority over key decisions and accountability for their outcomes have become clearer

Jo-Ellen Pozner, a professor at Santa Clara University, has also pointed out that new titles need performance goals and measurement/compensation systems to match. Beyond counting how many managers got cut, we need to look at how the restructured organization actually works.

References & Further Reading

• Jack Dorsey & Roelof Botha, “From Hierarchy to Intelligence”, Block, 2026. : This is the organizational operating vision put forward by Block’s leadership.
• “Goodbye, middle managers. Hello, ‘player-coaches’ and ‘org leads’”, Business Insider, 2026. : An article covering both companies’ restructuring and expert commentary on it.
• Block. Q4 2025 Shareholder Letter, 2026.
• Korn Ferry. Workforce 2025: Power Shifts.
• “Beyond the Holacracy Hype”, Ethan Bernstein, John Bunch, Niko Canner & Michael Y. Lee, Harvard Business Review, 2016. : An academic analysis of Zappos’ holacracy experiment. It offers a balanced look at both the promise and the limits of self-organization.
• Gartner. Predictions on IT and Organizational Change Beyond 2025, 2024.
• Business Insider. Meta Reality Labs’ Pilot Restructuring, 2026.

The author is Oswarld (Kwangseob Ahn). Current roles: Adjunct Professor at Sejong University, Strategy Consultant at INLEVEL9. Career, research, books, and recent work are kept current on the About page. Latest · July 2026: HEMA-2: A Consolidation-Aware Tri-Memory Architecture with Multi-Channel Scheduling for Lifelong Conversational AI.

Footnotes

1. Intelligence Layer: In Block’s essay, this refers to an AI function that combines features like payments or lending — based on company and customer information — to make proposals to customers. It connects to models that track internal company conditions, but the term isn’t limited to information-relay tasks. ↩

2. DRI (Directly Responsible Individual): Someone accountable for the outcome of a specific initiative. In Block’s plan, this person takes on problems that span multiple teams and coordinates the resources needed to solve them. How long someone holds the role can vary by task. ↩

3. Holacracy: An organizational operating model that distributes authority across roles and team units called “circles” according to a fixed set of rules. The responsibilities and authority attached to each role are continuously adjusted. ↩