Organizational Decision-Making 2026

We can build anything.
We can't decide what not to.

AI made building cheap. So "that looks good too," "this looks good too" multiply endlessly. The scarce resource is no longer implementation — it's the judgment of what NOT to build, what to stop, what to decide, and what to commit to. Yet human cognition is wired to add and to overlook what could be removed. This is the real shape of a new organizational debt — read its lineage and current state, with citations.

8 experiments
demonstrating that people choose to add over subtract (Nature 2021)
66%→56%
pick the most-featured product in store, but the simpler one after use (Feature Fatigue, JMR 2005)
+15%
Amazon's delayering target for the IC-to-manager ratio (by Q1 2025)
2015→2026
Bezos's reversibility frame, re-cited by Deloitte as a current AI-era recommendation
↓ SCROLL
01 / The Thesis

The scarce thing moved
from "building" to "deciding."

Agrawal, Gans & Goldfarb's Prediction Machines (2018) read AI through economics, not technology. What AI cheapens is "prediction" (≈ building); its complement — judgment — becomes scarce and expensive. As the marginal cost of building approaches zero, an organization's edge moves to the layer that decides what to build, and what not to.

Prediction Machines / 2018

Cheap prediction ⇒ scarce judgment

Recasts AI's output as a cheap commodity, "prediction." Judgment is its complement. Susan Athey: "what strategists and managers really need to know about the AI revolution." A book explicitly aimed at organizational decision-making.
Agrawal, Gans & Goldfarb — HBR Press
Field observation / 2026

Only "additions" reach the agenda

You can build anything now, so "that looks good too" multiplies without end. With no one empowered to say "stop," the product bloats and time drains away. This is a layer distinct from psychological safety.
An observation from building with AI
Taste discourse / 2026

"Choosing & cutting" is the new bottleneck

In the vibe-coding / agents era, building goes free and scarcity shifts to "the taste to choose, cut, and judge." A vernacular take on Prediction Machines. Note: this comes from practitioner blogs, not peer-reviewed work.
"Taste is the new bottleneck" et al. (blog)
02 / The Headwind

Cognition is the headwind.
People overlook subtraction.

"Can't decide" is not weak will. It's structure. From individual cognition to group aggregation, four demonstrated mechanisms push an organization toward addition whenever it is left unmanaged.

Nature / 2021

Additive bias

Verbatim: "People systematically default to searching for additive transformations, and consequently overlook subtractive transformations." Shown across 8 experiments. The authors connect it directly to overburdened schedules and institutional red tape. Independently replicated in 2025.
Adams, Converse, Hales & Klotz
Kyklos / 1966

Tyranny of small decisions

Coined by Kahn. A sum of individually small, sensible decisions yields a large outcome that is neither optimal nor desired. Each choice is a "vote" assuming the alternative stays available — keep adding, and complexity no one chose takes up residence.
Alfred E. Kahn — The Tyranny of Small Decisions
JMR / 2005

Feature fatigue

Before use, people overweight capability and underweight usability → they choose overly complex products. In store 66% pick the most-featured model; after use 56% prefer the simpler one. Too many features lower customer lifetime value.
Thompson, Hamilton & Rust
O'Reilly / 2018

Build trap

Melissa Perri: being stuck cranking out features (outputs) on schedule rather than customer value (outcomes). The cure is a culture shift from output to outcome. The organizational face of the "what to build" problem.
Escaping the Build Trap

How to read this (avoid over-claiming)

The additive-bias experiments measured individual cognition. Extending it to organizations is the authors' own reasonable inference (red tape / bureaucracy), not direct organizational evidence. A 2024 replication shows dependence on task, culture and age, weakening "universal default" — the bias reproduces on average, but isn't "always." Kahn's original mechanism is also a market aggregation across many consumers, a different category from a single organization. Tying these studies together as "organizational decision debt" is an interpretation, not a claim the originals make themselves.

→ Read it as "left unmanaged, there is structural pressure toward adding," not "people always add."

03 / Prior Art

The building blocks
already exist.

When you think through decision-making in the AI era, a set of well-tested prior concepts is already there to lean on. Below is the lineage — from the cognitive bias to codified decision rights.

Concept
Proposer / Year
Role
Subtraction neglect / Additive biasthe default to add over subtract
Adams, Converse, Hales & Klotz / 2021
★Cognitive engine; left alone, it always tilts to adding
Tyranny of Small Decisionssmall choices, big unchosen outcome
Alfred E. Kahn / 1966
★Group aggregation; complexity no one chose
Feature Fatiguetoo-many-features regret
Thompson, Hamilton & Rust / 2005
Proof that "adding" is locally optimal yet globally costly
Build Trap / Feature Factorybuilding becomes the goal
Perri 2018 / Cutler 2016
The organizational symptom (output dependence)
Type 1 / Type 2 Decisionsirreversible vs reversible
Jeff Bezos / 2015
★Core frame: calibrate speed by reversibility
Prediction Machinesjudgment as complement
Agrawal, Gans & Goldfarb / 2018
★Abundance anchor (judgment becomes scarce)
Organizational Debtdebt of missing scaffolding
Steve Blank / 2015
Org version of technical debt (adjacent; a different target)
RAPID / DRI / Disagree & commitcodified decision rights
Bain 2006 / Amazon & Grove
Repayment tools (who decides, and commits)
04 / The Core Frame

Two prescriptions:
calibrate speed, stop adding.

The core remedy for decision debt converges on Bezos's reversibility frame. And the decisive point: this frame isn't a 2018 relic — in 2026 a tier-1 consultancy re-cites it as a current recommendation.

CORE FRAME 01 — Bezos 2015
deliberate on allsplit speed by reversibility

Irreversible (Type 1 / one-way door): decide methodically, slowly, with consultation. Reversible (Type 2 / two-way door): high-judgment individuals or small groups should decide quickly — not by committee or consensus.

"As organizations get larger, there seems to be a tendency to use the heavy-weight Type 1 decision-making process on most decisions, including many Type 2 decisions. The end result of this is slowness, unthoughtful risk aversion, failure to experiment sufficiently, and consequently diminished invention." — Amazon 2015 Letter (verified against the primary PDF)

CORE FRAME 02 — additive → subtractive
only additions on the agendaalways table a subtraction

Additive bias works by never searching for the subtractive option. So the fix is design, not willpower: structurally force "stop / reduce" options onto the agenda, and put "added vs. retired" side by side on the roadmap.

People most overlook subtraction when not cued to consider it, when they have only one shot, and under higher cognitive load (Nature 2021). I.e., the busier the org, the more it adds.

2026 reality check — Deloitte Global Human Capital Trends

Deloitte's 2026 edition (published 2026-03) cites Amazon's one-way / two-way door model as a current recommendation for matching decision speed to reversibility in the AI era, advising organizations to "classify choices and pre-assign owners, data, guardrails, and speed for each category." So this core frame hasn't aged out — it is being re-requested for the AI era.
Note: Deloitte uses "one-way/two-way door" / "Amazon's 2015 shareholder letter," not "Type 1/Type 2" or "Bezos." The Type 1/2 label belongs to the Bezos letter.

05 / Org Design Now

Structure manufactures
"hard to decide."

Decision debt also lives in structure. More middle layers breed the sense that "the decision is made elsewhere," eroding ownership; annual-locked budgets can't track changing conditions. Here are 2025-26 cases, with hedges attached.

Amazon / Jassy 2024-25

Delayering — fewer layers, decisions to the front line

Jassy pledged to raise the ratio of ICs (front line) to managers by at least 15% by the end of Q1 2025 (target set 2024-09, reported achieved by 2025-03). The stated aim: undo the state where "owners of initiatives feel they shouldn't make recommendations because the decision will be made elsewhere," and restore ownership.
"flatten our organization and have fewer layers"
McKinsey / correlational

Nimble reallocation of capital

The most active reallocators — the top third shifted an average of 56% of capital across business units over 15 years — earned ~30% higher annual TRS than the bottom third. A sign of the edge of nimbleness over static, annual-locked allocation — but correlation, not causation.
~1,600 US companies / 1990–2005

How to read Amazon's number

The 15% is Amazon's self-report, with no external audit. The true cause is contested: analysts point to $2.1–3.6B in manager-pay savings, while Jassy frames it as "not really financially driven, not even really AI-driven" — about culture, agility, ownership. This note treats the stated rationale, not a verdict on the real cause.

→ Case evidence that fewer layers can speed decisions — not causal proof of effect.

How to read McKinsey's number

56% / 30% is correlational, as McKinsey itself frames it. Reverse causation is plausible (higher performers hold more cash, so they can reallocate more). The data is 1990–2005 — over 20 years old. The adjacent "10% vs 6%" / "10.8% vs 2.5% CAGR" figures often quoted nearby could not be verified here and are not used (only the 56%/30%/15-year version is).

→ Nimble reallocation is "promising but not causally established." Use it to prompt a budget rethink, no more.

06 / Decision Rights

Put "who decides"
on paper.

Most organizations that can't decide have ambiguous decision rights. The way out is to codify who decides, who inputs, and that once decided everyone commits. Three of the most canonical mechanisms.

Bain / 2006

RAPID

Five roles (Recommend / Agree / Perform / Input / Decide). Verbatim: "the Decider makes the final decision and commits the organization to action. Ideally, there should only be one decider for each decision." Stated purpose: transparency on accountability, reduced ambiguity.
Rogers & Blenko, "Who Has the D?" (HBR)
Amazon / Apple

Single-threaded owner / DRI

One dedicated owner per initiative. Unlike RACI's task-centric "Accountable," it is decision-centric — making "whose D is this?" unambiguous. The opposite of committee ownership (a decision belonging to no one).
STO / Directly Responsible Individual
Grove / Amazon

Disagree and commit

Disagree until it's decided. Once decided, everyone commits — even those who privately objected. Make not just "deciding" but "committing" part of the definition of done — the direct cure for decisions that never turn into action.
Disagree and commit

How far does the evidence go?

What could be verified here is the stated purpose (reducing ambiguity), no further. The phrasing that Bain asserts a direct causal link from decision accountability to organizational performance could not be supported and was rejected. Stated adoption is widespread, but peer-reviewed or quasi-experimental evidence that it improves velocity or outcomes is thin.

→ "Codify the rights" is sensible, but the effect awaits proof. Don't equate adoption with improvement.

07 / Open Questions

What's solid,
what's still thin.

The prior concepts are all here. But the questions specific to the AI era still have thin-evidence areas. Here, honestly, is what the research found to be not yet certain.

Open questions left after verification
Where things stand now
Systematic, multi-company implementations of AI-agent-native org redesign?
The biggest gap. What's verified is Amazon's human-centric delayering (Jassy himself denies an AI cause) and 20-year-old McKinsey stats. Evidence for agent-native orgs did not surface.
Causal, recent quantification that nimble reallocation beats annual-locking?
McKinsey stops at correlation. Causal evidence since 2015 is lacking. The current data backing the Capital Debt thesis is thin.
Do RAPID/DRI/DACI really raise decision velocity once codified?
Peer-reviewed / quasi-experimental proof of effect is thin. Widely adopted, yet causation is unestablished.
08 / Key Messages

People remember "lines,"
not "frameworks."

Lines that land — for meetings, slides, or a single sentence.

AI made building free. The scarce resource now is the power to decide what NOT to build.
Indecision isn't weak will. People structurally choose to add and overlook subtraction (Nature 2021). So the cure is design, not willpower.
Reversible decisions: fast, and by one person. Committee and consensus are poison precisely for the decisions you can undo (Bezos 2015 / Deloitte 2026).
Don't stop at "decided." Decide, then commit (disagree and commit) — that is the decision.

Sources

Sources checked against primary documents. Correlation is flagged as correlation, and claims that couldn't be substantiated are not used as support. Numbers are kept identical across JA/EN.

More in this series — AI Era Insights
Boards & C-Suite
Boards and the C-Suite in the AI Era
Hiring Strategy
The Talent Blueprint for the AI Era
Education Strategy
Entrepreneurship Education for the AI Era