Managers Are Not Overhead: They Are Infrastructure

David Michelson · article · 2026-07-22 source ↗ Reader 🖥️

AI as Amplifier AI Labor Disruption Coordination Cost Enterprise AI Adoption & Readiness Hierarchy & Org Charts Human as Orchestrator of Agents Organizational Legibility Psychological Safety

Organizations flattening management layers on the assumption that AI has absorbed most of what managers used to do are making an unsupported productivity bet, because converging survey evidence shows managers have become the critical infrastructure layer that translates AI investment into realized team-level value, not a layer of overhead AI is now positioned to eliminate.

practical · argumentative — A short opinion essay structured as a rebuttal: it states an opposing narrative (flatten because AI absorbs manager work), then marshals two survey sources (Microsoft's 2026 Work Trend Index, LeadDev's Engineering Leadership Report 2026) against it, then draws a conceptual distinction (flattening vs. contraction) to reframe the diagnostic question for the next round of cuts.

Structure

Part	Role
Opening: the flattening narrative	states the opposing thesis being rebutted — layoffs have targeted managers, AI can now do scheduling/reviews/coordination/decision-support, so 'everyone is a manager now' and dedicated managers become less necessary
Microsoft 2026 Work Trend Index data	supplies the counter-evidence: organizational factors (culture, manager support) account for twice the AI impact of individual effort; manager modeling of AI use and psychological safety produce large measured lifts in employee AI value, trust, and readiness
'Frontier Professionals' finding	strengthens the counter-evidence by showing the manager effect is largest precisely among the most advanced AI users — the population flattening advocates most want to unleash
Anticipated counter-rebuttal and its answer	grants that the role could still contract once AI adoption normalizes, but argues that if contraction were happening we'd see early signs; instead LeadDev's 2026 Engineering Leadership Report shows scope, communication load, team count, and direct reports all increasing
Flattening vs. contraction distinction	the essay's conceptual pivot: reinterprets 'more teams, more reports' as evidence of flattening working as intended, not of the role disappearing — contraction would mean the role's scope itself shrinks, which the data doesn't show
Concession and reframed question	concedes that genuinely unhelpful bureaucratic layers should still be cut, but reframes the test as: are you cutting based on what managers used to do, or on the critical work they do now/next?
Closing restatement	names the overhead/infrastructure distinction explicitly and states the productivity-bet claim as the essay's conclusion

Key terms

overhead
a framing of managers as reducible administrative cost tied specifically to tasks (scheduling, reviews, coordination) that AI can now perform — the target framing the essay argues against
infrastructure
the load-bearing organizational layer that converts an investment (in AI) into realized value at the team level — necessary substrate rather than optional layer
flattening Organizational Legibility
removing management layers / widening span of control — explicitly distinguished from contraction; flattening can coexist with an expanding, not shrinking, managerial role
contraction
narrowly defined as the scope of the managerial role itself shrinking because AI and ICs absorb more of the work — a specific claim the author argues the data does not support, as opposed to flattening
Transformation Paradox Enterprise AI Adoption & Readiness
borrowed from Microsoft: the condition where employees move faster with AI than the organization around them adapts; managers are cast as the layer that resolves it

Propositions

explicitManagers have been disproportionate casualties of post-COVID-19 tech layoffs, with 'flattening the org' and 'reducing bureaucracy' becoming synonymous with thinning management layers that grew during 2021–2022 hiring sprees.conf: high
extendsManagers Are Struggling to Keep Up with the AI Productivity Boom · p1 — Both observe AI-driven pressure on the manager layer; the HBR piece frames it as managers struggling to keep up with accelerated execution (a strain claim), while this essay extends the picture by arguing that very strain is evidence managers are becoming more, not less, essential.
explicitcontestedThe prevailing logic for cutting managers holds that because AI can now automate scheduling, performance-review drafting, cross-team communication coordination, and decision support, fewer managers are necessary and individual contributors can absorb managerial judgment themselves.conf: high
contradictsThe founder of Postman says you have to kill your existing org chart · p6 — This essay's opening statement of the narrative it goes on to rebut is the same claim the Postman-founder essay asserts affirmatively ('Middle management layers are primarily coordination overhead that AI tools make unnecessary') — direct opposition on the same claim.
explicitOrganizational factors — culture, manager support, talent practices — account for roughly twice the reported AI impact of individual effort alone (Microsoft's 2026 Work Trend Index Annual Report).conf: medium
explicitcorroboratedWhen managers actively modeled AI use, employees reported a 17-point lift in reported AI value, a 22-point lift in critical thinking about AI use, and a 30-point lift in trust in agentic AI; when managers created psychological safety around experimentation, employees reported up to 20 points higher AI readiness/value and were 1.4x more likely to be high-frequency agentic-AI users.conf: medium
supportsHow tech workers actually feel about AI in 2026 | Annual AI sentiment survey (Noam Segal) · p5 — Microsoft's finding that manager AI-modeling behavior drives large employee AI-outcome lifts is a specific mechanism supporting the broader survey finding that manager effectiveness is the single largest actionable lever on employee wellbeing/outcomes.
extends ←The Hidden Realities of AI Adoption · p9 — Reframes the trust-lift Microsoft's Work Trend Index measures (manager AI-modeling -> +30pt trust in agentic AI) as something organizations should track prospectively as a leading indicator, not only interpret retrospectively as a correlational outcome of manager behavior.
supports ←What to Do When There's Too Much Work · p12 — Both identify psychological safety around AI experimentation/adoption as a high-leverage lever for organizational AI outcomes — this document as an assertion, the other with measured lift figures.
explicitThe positive effect of manager behavior on AI value is largest among 'Frontier Professionals' (16% of respondents who use agents for multi-step and multi-agent workflows), who are far more likely than others to report a manager who uses AI, sets quality standards, encourages experimentation, and rewards work redesign.conf: medium
explicitManagers are the organizational layer that resolves Microsoft's 'Transformation Paradox' — employees moving faster with AI than the organization around them — by translating leadership's AI strategy into team-level practice.conf: medium
supportsAI Won't Save the Kingdoms We Built · p7 — Both frame AI as valuable only through an existing coherent human layer — Cutler's 'AI is only genuinely useful as an amplifier of an already-coherent operating model' matches this essay's claim that managers are the layer translating AI strategy into team-level practice.
explicitIf management roles were already contracting under AI, early signs should be visible; instead LeadDev's Engineering Leadership Report 2026 (600 engineering leaders surveyed) shows 63% report increased scope/responsibility, 60% increased communication with teams/customers/stakeholders, 22% more teams reporting to them, and 29% more direct reports, alongside more hands-on technical involvement.conf: high
extendsEngineering Leadership Report 2026 · p5 — Both draw on the same LeadDev Engineering Leadership Report 2026 to document managers taking on more hands-on and higher-scope work; this essay adds further survey items (communication load, team count, direct reports) from the same report as additional evidence of role expansion rather than contraction.
explicitFlattening (fewer managers, wider spans of control) is conceptually distinct from contraction (the scope of the managerial role itself shrinking); more teams and more direct reports per manager is what flattening produces, not evidence that the role is disappearing.conf: high
supportsAI Won't Save the Kingdoms We Built · p9 — Both are skeptical that flattening simply removes coordination need; Cutler argues the coordination burden migrates and resurfaces elsewhere, this essay argues flattening's visible effects (more scope, more reports) are evidence the underlying managerial work hasn't actually shrunk.
refinesThe engineering manager role is splitting in two · p8 — Both cite the same empirical flattening signals (Meta, Amazon, Gartner data); this essay refines the interpretation by arguing the signals indicate flattening rather than contraction of the underlying role.
explicitNone of this implies organizations should stop scrutinizing reporting structures or removing genuinely unhelpful bureaucratic layers; the real test for the next round of cuts is whether reductions are based on what managers used to do versus the critical work they are doing now or will need to do next.conf: high
explicitFlattening on the assumption that AI will drive its own adoption, or that value will emerge from unguided individual effort, is a productivity bet that the evidence does not support.conf: medium

Arguments

Organizations cutting managers because AI has absorbed much of their traditional workload are making an unsupported productivity bet; managers should instead be understood and preserved as the infrastructure layer that lets AI investment convert into realized value. strength: medium

Explicit premises:

Manager behavior (modeling AI use, creating psychological safety) is strongly associated with employees' AI value, trust, and readiness (Microsoft Work Trend Index)
This effect is largest among the most advanced AI users, the population organizations most want to accelerate
If the managerial role were genuinely contracting under AI, early signs should already be visible; instead LeadDev's 2026 survey shows managers' scope, communication load, team count, and direct reports have all increased

Implicit premises:

The correlations reported by Microsoft (manager AI-modeling behavior <-> employee AI outcomes) reflect managers causing the improved outcomes, rather than reverse causation or a shared confound (e.g., already high-functioning, high-trust teams both adopt AI well and have managers who model its use)
The tasks AI can now automate (scheduling, drafting reviews, coordinating communication) are separable from the 'critical work' that actually determines whether AI investment pays off, so automating the former does not diminish the case for the latter
Rising scope, communication load, and direct-report counts documented by LeadDev represent additional value-generating managerial work rather than a sign of an already-strained role being asked to do more without more time or support

Inferences:

Never draws the inference that if flattening keeps widening spans of control, managers' capacity to perform the very behaviors credited with driving AI value (individualized modeling, psychological safety, quality standard-setting) may itself erode — the argument for 'managers matter more' doesn't reckon with flattening simultaneously reducing the time managers have per report
Never questions whether Microsoft (an AI-tooling vendor) and the Postman-sponsored LeadDev survey have a shared commercial interest in findings that argue for preserving manager seats who are also the ones who purchase and roll out enterprise AI tools

Narratives

The layoff wave at Meta, Google, and Amazon, and the extreme case of '50 ICs reporting into one supervisor' at Meta's applied-AI engineering reorganization
→ anchors the flattening narrative in concrete, high-profile examples before the essay pivots to rebut the logic behind them

Questions addressed

✅ Is cutting managers because AI can perform their traditional tasks a sound basis for flattening?
No — the essay argues managers' value now lies in translating AI investment into team-level results, not in the administrative tasks AI can automate.
✅ Is the managerial role actually contracting under AI adoption?
No — LeadDev's 2026 data shows scope, communication, team count, and direct reports increasing rather than shrinking.
✅ Does manager behavior affect how successfully employees adopt and benefit from AI?
Yes, per Microsoft's Work Trend Index — manager modeling and psychological safety correlate with large lifts in employee AI value, trust, and readiness.
⭕ Are the widening scope, communication load, and report counts LeadDev documents a healthy sign of managers becoming more valuable, or a sign of an already-strained role being stretched further without added time or support?
Not addressed — the essay treats the increases only as evidence against contraction, without asking whether they are sustainable.
⭕ Will the managerial role eventually contract once AI adoption becomes the norm and managers no longer need to manage the change itself?
Explicitly left open: 'We don't know how this will play out yet.'

Gaps

illogicalThe Microsoft Work Trend Index correlations (managers who model AI use and create psychological safety <-> higher employee AI value/trust) are treated as showing managers cause the AI-value-realization effect, but the reported data doesn't rule out reverse causation or a shared confound — e.g., managers who already lead high-functioning, high-trust teams may both model AI use more and see better AI outcomes for reasons unrelated to the modeling itself. This is the same causal-inference gap the corpus has repeatedly flagged for similar manager-effectiveness survey findings.
Read next: Read against extract-lenny-noam-segal-tech-worker-ai-sentiment-2026#g1, which raises an identical causation-vs-selection concern about a structurally similar manager-effectiveness finding, for a methodologically stronger version of this claim (quasi-experimental studies of manager rotation/turnover on team AI outcomes).
incompleteThe essay never reconciles the tension its own evidence walks into: the LeadDev data it cites to show managers 'doing more' (more scope, more reports, more communication) is, elsewhere in this corpus, flagged as an unresolved structural contradiction — manager quality/attention is the single biggest lever on wellbeing, yet spans of control keep widening. If managers are stretched across more reports and more scope, their capacity to perform the very behaviors credited with driving AI value (individualized modeling, psychological safety, quality-standard setting) may itself erode, an inference the essay never draws.
Read next: Reads as a partial answer to the reconciliation gap named in extract-lenny-noam-segal-tech-worker-ai-sentiment-2026#g2 and extract-tech-workers-2026-workforce-splitting-in-two#g4 (orgs flattening while manager quality is the biggest wellbeing lever) and extract-postman-engineering-leadership-report-2026#g5 (the same report's own reassurance-vs-accelerating-flattening tension) — but it only asserts a side (invest in/preserve managers) rather than explaining how broader spans and deeper manager investment can coexist.
uninformedBoth empirical sources cited — Microsoft's own Work Trend Index and the Postman-sponsored LeadDev Engineering Leadership Report — are produced or sponsored by parties with a commercial stake in enterprise AI seat and tool adoption (Microsoft Copilot; Postman's API platform, whose customers are the same engineering-leadership audience). The essay treats both as neutral evidence for 'invest in managers' without noting that a report emphasizing manager-mediated AI success also implicitly argues for retaining the manager seats who buy, roll out, and champion AI tooling.
Read next: Compare to extract-postman-engineering-leadership-report-2026#g3, which flags the identical sponsored-study pattern for this same LeadDev report.
unaddressedThe essay rebuts the 'AI replaces manager tasks so cut managers' narrative in general terms ('managers are infrastructure') but never engages the sharper, competing claim already in circulation that AI eliminates specifically the *coordination* function managers historically performed (not managerial value generally) by letting ICs and PMs ship directly — a claim that argues for redefining, not simply preserving, the role.
Read next: Read against extract-kill-your-org-chart-for-ai (management's purpose becomes developing judgment, not coordination) and extract-engineering-manager-role-splitting-in-two (the TLM/multi-team-EM fork) for competing accounts of which manager functions AI actually obsoletes versus elevates.

Closes gaps elsewhere

How tech workers actually feel about AI in 2026 | Annual AI sentiment survey (Noam Segal) · g2 — Argues directly for one side of the named reconciliation: organizations should weight the emerging evidence that manager behavior drives AI value over the impulse to keep widening spans of control. This only partially closes the gap — it doesn't specify a concrete mechanism for reconciling broader spans with deeper manager investment (see this extraction's own g2 above).
Engineering Leadership Report 2026 · g5 — Supplies the interpretive frame the Postman/LeadDev report itself lacks for its internal tension between 'adapting, not dismantling' reassurance and accelerating upper-management flattening: the flattening-vs-contraction distinction reframes rising scope/reports as compatible with (not contradicting) an argument that managers remain critical.
How tech workers are feeling in 2026: a workforce splitting in two · g4 — Same partial reconciliation as above: proposes that the correct response to 'manager quality is the biggest wellbeing lever, yet orgs keep flattening' is to stop treating managers as automatable overhead, without addressing whether flattened spans leave managers enough capacity to actually deliver that quality.