Section 01The 2010s Leadership Industry Trained Managers for the Wrong Workplace
Between roughly 2010 and 2022, the corporate leadership-development industry shipped substantially the same curriculum to substantially the same managers in substantially the same companies. Servant leadership. Growth mindset. Psychological safety. OKRs. Radical candor. The vocabulary varied; the operating assumptions did not. Most of that curriculum is still useful as ethics. Almost none of it addresses what AI augmentation does to the manager's actual job.
The assumptions buried inside the 2010s curriculum were that an individual contributor's output was bounded by their training and effort; that a manager held a meaningful information advantage over their reports; that the unit of performance management was an annual review cycle; that skills aged on a five-to-ten-year horizon; and that reorganizations were episodic events to recover from rather than steady-state weather. By 2026, every one of those assumptions is either weakened or broken. The leadership curriculum has not yet caught up.
This piece is not an argument that the older virtues stopped mattering. Trust, candor, equity, psychological safety — these remain ethical baselines, and the firms that abandoned them in pursuit of "efficiency" produced the most visible failures of the last three years. The argument is narrower and more operational: the manager's operating model — what they actually do with their hours, what they are measured on, what high performance looks like in the role — has shifted, and most managers were trained for a job that no longer exists in the form they were trained for it.
What this piece is not
Not a defense of "AI will replace middle managers." That framing is mostly marketing. The first-line manager role is changing more in its content than in its headcount, and the senior-leader role is changing more in its decision-architecture than in its existence. The interesting question is what the new work looks like.
Section 02What's Actually Changed About the Manager's Job
Five concrete shifts, each documented in the empirical literature or in the disclosed practices of identifiable firms. None of them is speculative; the speculation is about how managers and the institutions that train them respond.
1. The IC output curve has shifted, hard
The top-quartile individual contributor in a knowledge-work function, equipped with current AI tools and using them well, is producing somewhere between 2× and 10× their 2022 output, depending on the function. The GitHub developer-productivity studies, the McKinsey 2026 update on generative-AI work effects, and the Microsoft Work Trend Index 2024 all converge on the same shape: the distribution of output has gotten longer-tailed, and the gap between the median and top-quartile worker is wider than at any point in the measured history of the function [1][2][3]. The implication for managers is that "managing the average" is now badly miscalibrated for a team where the variance in productive output is several times what it used to be.
2. The manager's information advantage has narrowed
For most of the last fifty years, the manager held an asymmetric information advantage over their reports: more institutional context, more cross-team visibility, more domain experience. The junior contributor with AI tools available has closed a non-trivial fraction of that gap. A second-year analyst with a strong AI workflow can produce a synthesis of an unfamiliar market or codebase that, in 2022, would have required a senior. This does not eliminate the value of judgment, taste, or relationship capital — it relocates the manager's advantage from "I know more facts" to "I know which facts matter and what to do about them."
3. The performance-evaluation problem has gotten harder
If a junior worker's apparent output is 3× their 2022 baseline, how much of that is the worker and how much is the tooling? If a senior worker's apparent output is 1.4× their 2022 baseline, are they underperforming or are they doing harder work that resists augmentation? Activity metrics, which were a poor proxy for output before, are now substantially worse. The full treatment of this is in the Productivity Theatre pillar; here it is enough to flag that "managing performance" is the management activity whose ground-truth has shifted most.
4. The skill-currency horizon has compressed
The training-and-development industry built its curricula around a "five-year skill" — a competency a worker could acquire and rely on for half a decade before refresh. By 2024, the effective half-life of a discrete tool-skill in many knowledge-work functions is closer to eighteen to twenty-four months. The manager's job of "developing your people" now includes a deliverable it did not used to: keeping the team's skill currency current at the pace the labor market is repricing it. The full case for treating skill currency as a measurable manager output is in the Reskilling for Real pillar.
5. Reorganization frequency has roughly doubled
The McKinsey organizational-design dataset shows that the median large enterprise underwent meaningful organizational restructuring about once every 28–32 months in the 2012–2018 window. The same metric for 2022–2025 is closer to 14–18 months [4]. A manager who started in 2018 expected to operate inside a stable org chart for two or three years between reorganizations. A manager starting in 2024 should expect a reorganization within their first 18 months and another within the following 18. The "operate inside the chart" skillset is now subordinate to the "lead a team through structural change" skillset.
Section 03The Five Components of the New Operating Model
The shifts above do not require managers to abandon what they learned. They require managers to reweight what they do with their hours. The pattern that the highest-performing teams in 2024–2026 share is not a personality — it is an allocation of management attention. Five components.
(1) From activity-watcher to outcome-architect
The first and largest shift. The manager's central question moves from "is my team active enough?" to "is my team producing the right outcomes, and is the system around them designed to make those outcomes producible?" The full operational treatment of this — outcome metrics by function, the failure modes of activity-based evaluation, the 2×2 framework for thinking about productivity instrumentation — is in the Productivity Theatre pillar. The single test for whether a manager has internalized the shift: can they describe, in concrete numbers, what success for their team looks like at the next quarter boundary, without referencing hours, meetings, tickets, or messages?
(2) From subject-matter-expert to capability-translator
For decades, the implicit promotion ladder in most knowledge-work functions ran through technical mastery: the best engineer becomes the engineering manager, the best designer becomes the design lead, the best analyst becomes the analytics lead. The mastery-as-credential model is breaking down, not because mastery has stopped mattering, but because the technical surface of every function is expanding faster than any individual manager can track in detail.
The replacement model: the manager does not need to be the strongest practitioner. They need to know what good output looks like, to be able to evaluate it without producing it themselves, and to translate between functions whose languages have diverged. Camille Fournier, in The Manager's Path, makes the related point that the engineering manager's value is not technical depth but the ability to make tradeoffs visible to the people who can make them [5]. The augmented workplace generalizes the point to every function.
The role of the manager is not to be the smartest person in the room. The role of the manager is to multiply the leverage of the people in the room. Camille Fournier, The Manager's Path (O'Reilly, 2017)
(3) From scheduler to skill-currency steward
The traditional manager-of-development role assumed that workers had a five-year skill horizon and an annual development conversation. The augmented workplace has compressed both. The manager who treats their team's skill currency as a passive expectation — "people will figure it out, that's their job" — is shipping a team whose currency degrades faster than the labor market reprices it.
The shift is to treat skill currency as a measurable manager deliverable. Concretely: a quarterly review of which tools and methods the team uses, where the team sits relative to the function's frontier, what the gap is, and what investment closes it. Adam Grant's Hidden Potential develops the underlying argument that developmental capacity is the most underestimated variable in talent decisions [6]; the operational consequence in 2026 is that the manager who systematically develops their team outperforms the manager who hires for finished-product skills and watches them age.
(4) From rater to coach
The annual-review-as-judgment model assumed that a once-yearly compression of a worker's performance into a rating was a workable instrument for development and pay decisions. The model was never very good. In an environment where output volatility is higher, where the worker's tools change quarterly, and where the manager's information advantage has narrowed, the model is worse.
The shift is toward continuous coaching as the manager's central performance work, with rating as a necessary but downstream administrative function. Kim Scott's Radical Candor is the most-cited articulation of the principle — direct, frequent, care-personally-challenge-directly feedback as the unit of performance development [7]. The augmented workplace makes the rating decision harder and the coaching decision more important; reallocating manager hours toward coaching is the single most consistent practice across teams that are visibly performing in 2024–2026.
(5) From individual-mentor to sponsorship-architect
Mentorship — informal advice and guidance — is the version of developmental support that scales poorly and that managers default to because it is low-commitment. Sponsorship — actively spending the manager's political capital to put a team member in front of opportunities, decisions, and senior audiences — is the version that produces measurable career outcomes, and is the version managers most consistently underdeliver. The full treatment of the gap is in the Mentorship Mismatch pillar.
The shift is to treat sponsorship as a managerial deliverable with a count attached. How many times this quarter did you put a direct report in front of a decision-maker who could change their trajectory? How many cross-functional opportunities did you place them into? Sponsorship as a measurable output, not an informal nice-to-have, is what separates the manager-as-career-multiplier from the manager-as-task-router.
Section 04What This Means for First-Line Managers
The biggest single shift falls on the first-line manager — the manager of individual contributors, not of other managers. The first-line manager is closest to where the augmentation is happening, the furthest from the strategic conversations that justify their role to senior leadership, and the most exposed to the calendar pressure that the older operating model imposed. Three concrete moves matter most.
Fewer status-asking meetings, more blocker-removing meetings. The 1:1 conversation that opens with "what are you working on?" is the artifact of an operating model where the manager needed status information to do their job. In the augmented workplace, status is largely visible in the work artifacts themselves. The 1:1 that opens with "what's blocking you, and what would unblock you?" is the artifact of an operating model where the manager's value is in removing structural friction the IC cannot remove themselves. The shift is small in vocabulary, large in effect.
Less people-managing, more system-designing. The first-line manager's leverage in 2026 is less in the individual conversations they have and more in the team-level systems they design: how the team's work intake works, how decisions get made, how outputs get reviewed, how onboarding works for the next hire. Andy Grove's High Output Management was making this argument in 1983 — that the manager's output is the output of their organization, and the leverage is in the systems, not the individual conversations [8]. The argument is more true now, not less.
A manager's output is the output of the organizational units under his or her supervision or influence. Managerial productivity — that is, the output of a manager per unit of time worked — can be increased in three ways: by increasing the rate with which a manager performs his activities, by increasing the leverage of his activities, and by changing the mix of his activities to those with higher leverage. Andy Grove, High Output Management (Random House, 1983)
Defend the unstructured hours. The first-line manager whose calendar is back-to-back meetings has, by construction, no time to do the system-design work that produces leverage. The most-cited finding in the Microsoft 2023 and 2024 Work Trend Index reports is that managers in high-meeting environments report substantially lower team output, substantially higher burnout, and substantially worse retention — and the gap has widened since AI tools made the IC's day more compressible than the manager's [2]. The unstructured manager hour is not a luxury. It is the input to the highest-leverage manager output.
One concrete reallocation
If your weekly calendar shows more than 60% of working hours in scheduled meetings, the new operating model says you are over-scheduled by a factor that is structurally preventing the leverage work. The fix is not "be more efficient in meetings." The fix is fewer meetings. This is uncomfortable to act on. It is the most consistent operational characteristic of high-output managers in 2024–2026.
Section 05What This Means for Senior Leaders
The senior leader's job — the manager of managers, the function head, the executive — is changing in different ways. The first-line-manager problem is mostly about reallocating hours within a known role. The senior-leader problem is mostly about choosing what stops being a job, what becomes a job, and what the surrounding system has to look like for the augmented team to produce its expected output. Three areas absorb most of the change.
Org design for augmented teams
The team sizes, reporting structures, and span-of-control assumptions that the corporate org-design playbook codified in the 2000s were tuned to an output curve that no longer exists. If a top-quartile IC is producing 3–5× their 2022 output, the team structure that produces the firm's outcomes is not the team structure of 2022 with each role filled by a more-productive person — it is, in some cases, a smaller team with different role definitions; in others, the same team with materially different work allocated to it. Deciding which is which is senior-leader work, not first-line-manager work, and most firms are doing it badly because the decision is being delegated rather than owned.
Outcome metrics that survive scrutiny
The function head who cannot articulate, in concrete numbers, what their function is producing for the firm — and the senior leader who delegates that articulation downward and then accepts whatever the team self-reports — is the structural cause of most "productivity theatre" in the firm. The senior-leader deliverable here is a small set of outcome metrics that survive scrutiny: that connect to revenue, retention, quality, or customer outcomes; that cannot be gamed without producing the underlying outcome; and that the function defends in public. The DORA framework for engineering is the canonical example; equivalents exist or are constructable for every major function [9].
Deciding what stops being a job
The hardest senior-leader decision in 2024–2026 is not "what new roles do we create?" but "what existing roles stop existing in their current form?" The implicit deal between the firm and the worker — that the role you were hired for is broadly the role you'll do for several years — is under more strain than at any point since the 1990s reorganizations. Senior leaders who handle this with transparency, with notice, with serious investment in transition, build trust that pays back across the next decade. Senior leaders who handle it with the late-Friday-announcement pattern erode the trust capital that the augmented workplace depends on.
The senior-leader role in 2026 is less the choreographer of the meeting calendar and more the architect of the decisions that the firm will live with for years. Less performance art, more decision capital.
Section 06Three Patterns That Don't Survive Contact With the Augmented Workplace
Three management patterns that were viable, or at least tolerated, in the 2010s and that observably break in 2024–2026. Each has identifiable failure modes that show up in retention, output, and team eNPS within two to four quarters.
(a) The "best engineer in the room" trap
The technical manager who maintains their identity as the strongest practitioner — the senior engineer who became the engineering manager but still writes the most code, the design lead who still ships the most pixels — is the most over-represented failure mode in the engineering-management literature, and it has gotten worse in the augmented workplace. The manager's individual output now competes against AI-augmented ICs whose output volume the manager cannot match while doing the manager job. The pattern is unsustainable, the team underdevelops because the manager will not let go, and the manager burns out trying to do both jobs. The exit is to do the work Camille Fournier and Andy Grove both describe: choose. The choice has to be the manager job, or back to IC. Trying to be both is the most reliable way to fail at both.
(b) The "model the grind" pattern
The leadership-by-stamina archetype — long hours, always-on availability, performative weekend Slack messages, the implicit message that the worker who doesn't match the pattern is uncommitted — was a tolerated dysfunction in the 2010s. It is corrosive in 2024–2026. The augmented workplace produces enough output volume that the grind theater is no longer the variable that distinguishes high-performing teams from low-performing ones — it is, increasingly, the variable that distinguishes teams with high turnover from teams with low turnover. The full case for why the grind pattern is empirically losing is in the Burnout Decade pillar. The operational consequence for leaders: stop modeling it, and stop tolerating it from the managers below you.
(c) The "I don't need to learn the tools" manager
The manager who has decided that AI tool fluency is the IC's job, not theirs, is leading a team using instruments they cannot evaluate. They cannot judge whether their team is using the tools well, badly, or at all. They cannot calibrate what good output looks like. They cannot tell whether a junior's apparent productivity is the worker, the tool, or the prompt. They cannot model the behavior they need the team to adopt. In every team-level survey of "what predicts a high-performing AI-augmented team," manager fluency with the tools the team uses is among the top three variables. The fix is not to become the team's best prompt engineer; the fix is to use the tools enough to be able to evaluate the work the tools produce.
Section 07For Managers in the First 90 Days of Adapting
Concrete moves for a manager who has read the above and wants to act inside a single quarter. None of these require headcount, budget, or org-design changes. All of them are inside the manager's authority over their own calendar and their team's operating norms.
- Spend five hours a week as an IC, using the same AI tools your team uses. Not as performance, not as theater — as actual production work, ideally on something that ships. This is the single highest-ROI calibration exercise available to a manager in 2026. You will discover, within the first month, what the tools are good at, where they fail, what your team is actually doing all day, and what "good output" looks like in your function's current state. Without this, you are managing a system you cannot read.
- Rewrite your team's success metrics in outcome language. Pull whatever you currently use to describe the team's performance — to your manager, in 1:1s, in the team's own goal-setting. If it contains words like "velocity," "throughput," "hours," "tickets," or "activity," it is activity language. Translate every metric into outcome language: what does the customer get, what does the business get, what changes for the user. If the translation is impossible, the metric is probably not measuring what you thought it was.
- Audit your calendar against the five components. Print last week's calendar. For each meeting, ask: which of the five components (outcome architect, capability translator, skill-currency steward, coach, sponsorship architect) did this hour advance? If the answer is "none," that hour is a candidate for deletion. The first audit will be uncomfortable; the third will be habit-forming.
- Identify one structural blocker for your team and ship its removal inside 30 days. Not a meeting about it. The removal. The blocker that produces the most visible friction for the most members of your team — the broken handoff, the missing access, the meeting that should not exist, the approval loop that adds nothing — and act on it as if your performance review depended on it. In the new operating model, it does.
- Pick one direct report to actively sponsor this quarter. Specifically. By name. With a target — a meeting they should be in, a project they should lead, a senior they should be visible to. Sponsorship as a measurable count, not a vibe. Repeat next quarter with a different name. Two years in, you have built the trust capital that the augmented workplace runs on.
The manager's job has not gotten smaller. It has gotten different — and the managers who treat the change as a calendar problem will keep finding that their calendar is the symptom, not the cause.
The 2010s leadership-development industry will catch up. It is being rewritten as we speak, in books and curricula that will land in 2027 and 2028. The managers who do the rewriting themselves, in their own calendar and their own team's operating norms, in 2026, will be the ones the industry studies. The operating model is available. The question is who picks it up first.
