The Flight Manual for Zero Human Companies

By Camiel Notermans × Polsia  ·  Weekend Debrief  ·  April 4, 2026

Executive Summary

Four humans are circling the Moon right now — made possible not by eliminating human judgment, but by encoding it so precisely that failure becomes improbable. The aviation checklist was born from catastrophe in 1935. The WHO surgical checklist cut death rates by 47% in 2009. Both prove the same thesis: complexity doesn't demand fewer humans — it demands better protocols.

The Zero Human Company is not a headcount reduction exercise. It is a protocol design exercise. The companies that grasp this distinction first will be the ones still operating in 2030. Most boards are debating AI adoption. The relevant question is AI governance: what gets decided autonomously, what gets escalated, and who is accountable when the system drifts. The flight manual is not a constraint on autonomy. It is what makes autonomy possible at scale.

The Mission That Proves the Thesis

As of this morning, NASA astronauts Commander Reid Wiseman, Pilot Victor Glover, Mission Specialist Christina Koch, and CSA astronaut Jeremy Hansen are 72 hours into the Artemis II mission — humanity's first crewed flight beyond Earth orbit since Apollo 17 in 1972. They named their spacecraft Integrity.

The translunar injection burn completed on April 2 at 7:49 PM EDT. Flawless, per NASA flight director Rick Henfling. "Human beings have left Earth orbit," said NASA acting associate administrator Lori Glaze. For the first time in more than fifty years.

Here is what made that burn possible: not the courage of the crew, though that is real. Not the computing power of the Orion spacecraft, though that is real. What made that burn possible is a document — a precisely sequenced, battle-tested, obsessively revised flight manual that tells four humans and thousands of ground controllers exactly what to do, in what order, under what conditions, and when to stop. The system that carried those four people 250,000 miles from Earth runs on protocol.

October 30, 1935: The Day the Checklist Was Invented

Ninety-one years before Artemis II, the most advanced aircraft in the world killed two of the best test pilots in the US Army Air Corps.

The Boeing Model 299 — what would become the B-17 Flying Fortress — was a marvel of engineering. Four engines. Far greater range and payload than anything else in existence. Boeing had a government contract within reach. The Army Air Corps sent their best: Major Ployer Peter Hill, chief of the Wright Field Flying Branch, with over 60 aircraft types in his logbook.

The Model 299 lifted off, climbed to 300 feet, stalled, and crashed. Investigation found the cause: the gust locks — controls that hold the aircraft steady on the ground — had not been released before takeoff. A procedure so basic it had never needed to be written down. Until now it had been flown by pilots who knew it cold. The Model 299 was too complex for any single pilot's memory to hold.

Boeing was nearly destroyed. The company survived by inventing the checklist. They flew 18 B-17s for 1.8 million hours without incident. Nearly 13,000 were eventually built. The checklist spread to all military aviation, then commercial aviation, then everywhere. "The loss of Hill and Tower," wrote aviation historian Doyle, "was directly responsible for the creation of the modern written checklist used by pilots to this day."

The lesson was not that the pilots were incompetent. It was that the system had outgrown unaided human memory. The response was not to simplify the system — it was to encode it.

Commercial Aviation: What Systematic Protocol Produces

In 2025, commercial aviation carried just under five billion passengers on more than 38 million flights. The IATA 2025 Annual Safety Report recorded a historic milestone: zero Loss of Control Inflight accidents — historically the leading cause of aviation fatalities — across the entire year.

Commercial aviation is, by any measure, the safest mode of mass transportation ever operated. Its safety is not the result of cautious engineering alone. It is the product of a relentless commitment to Standard Operating Procedures: precise, written, role-specific protocols that define what every crew member does, in what sequence, under every foreseeable condition, and — critically — what to do when something unforeseen occurs.

Captains and first officers on a modern commercial flight do not operate from experience and instinct alone. They operate from a manual. The manual does not diminish their expertise — it channels it. The manual defines what is delegated to automation, what requires human confirmation, and what requires human override. The aircraft does not fly itself. The crew does not fly it by gut feel. The manual makes the partnership work.

That partnership — human authority structured by explicit protocol, with autonomous systems executing within defined boundaries — is the architecture Airbus called "fly-by-wire with authority protection." It is also the architecture every serious autonomous company will need to build.

The Surgeon's Checklist: Same Logic, Different Stakes

In 2007, Atul Gawande, surgeon and staff writer for The New Yorker, joined a WHO initiative to address preventable surgical deaths. He drew the same conclusion the Boeing engineers had drawn in 1935: the problem was not inadequate surgeons. The problem was systems too complex for reliable unaided execution.

His team designed the WHO Surgical Safety Checklist: 19 items, executable in under two minutes. Three checkpoints — before anesthesia, before incision, before the patient leaves the room. Team introductions. Allergy confirmation. Equipment verification. Nothing exotic.

The results, published in the New England Journal of Medicine in 2009, were extraordinary: death rates fell from 1.5% to 0.8% — a 47% reduction. Complication rates fell from 11% to 7%, a 36% reduction. These results held across eight hospitals on six continents, in both high-income and low-income settings. The checklist was subsequently adopted in over 20 countries and called "the biggest clinical invention in thirty years" by The Independent.

Gawande's central argument was not that surgeons are fallible in some unusual way. It was that all human experts are fallible under the conditions of complexity, pressure, and cognitive load that define modern professional work. The checklist does not replace expertise. It protects expertise from the conditions that reliably degrade it.

The ZHC Thesis: The Manual Is the Moat

Three industries. Three different eras. The same structural insight.

When systems become complex enough that unaided human memory and real-time judgment cannot guarantee reliable outcomes, the answer is not to reduce complexity — and it is not, contrary to popular belief, to remove humans. The answer is to encode the system's operating logic with enough precision that the partnership between human and machine becomes structurally reliable.

This is the Zero Human Company thesis, properly understood.

The companies deploying AI autonomously at scale right now — Palantir with its AI-driven analytical workflows, Klarna with its near-elimination of traditional customer service headcount — did not simply purchase AI and launch it. They redesigned their operating procedures. They defined, explicitly, what the AI decides, what it escalates, what conditions trigger human review, and who is accountable when the system produces an unexpected outcome. They wrote a flight manual.

The boards that are currently debating AI adoption are asking the wrong question. The relevant question is: What is our operating doctrine for AI? Not "which tools do we buy" but "what are our mission rules" — the term NASA uses for the precise, pre-agreed decision protocols that govern every contingency a flight crew might face.

Without that doctrine, AI deployment produces one of two failure modes. Either the organization becomes over-dependent on autonomous systems whose failure modes are poorly understood, or it deploys AI with so many human approval layers that the efficiency gains evaporate entirely. Neither is a Zero Human Company. Both are expensive.

The Hard Truths

Protocol is not bureaucracy. The companies that treat AI governance as a compliance checkbox will fail to capture the productivity gains. The companies that treat it as competitive architecture will compound those gains quarterly.

The bottleneck is not AI capability. As of this quarter, AI capability substantially exceeds most organizations' ability to deploy it reliably. The constraint is operating doctrine — not the technology.

Most "AI transformation" programs are running without a flight manual. They have deployed AI tools. They have not defined authority limits, failure thresholds, escalation triggers, or accountability structures. They have handed the crew an aircraft with no checklist.

The companies writing their flight manuals now are building a structural moat. Operational doctrine, once embedded in systems and culture, is genuinely difficult to replicate. The first-mover advantage in autonomous AI governance is not the AI — it is the protocol.

The 30-Day Board Actions

Map current AI authority. Document every AI decision in your organization that currently executes without human review. That is your current flight manual. Most organizations will discover they do not have one — they have a collection of ad hoc deployments with no coherent doctrine.

Define three tiers of AI authority. Autonomous execution (no human required), supervised execution (human reviews exception flags), and escalated execution (human confirms before action). Assign every current AI deployment to a tier.

Establish failure protocol. For each autonomous AI workflow, define: what failure looks like, what the automatic response is, and who is accountable. NASA calls these "mission rules." Write yours.

Run a tabletop exercise. Simulate an AI system producing an unexpected outcome in one of your autonomous workflows. Does your team know what to do? Does a protocol exist? The answer, for most boards, will be clarifying.

The Mission Succeeds Because the Manual Was Written First

The Artemis II crew did not lift off and hope for the best. Every contingency was pre-decided. Every failure mode had a named response. Every AI-assisted system had a defined authority limit — a precise point at which human judgment would override automated execution. The manual was not written after the anomaly. It was written before the rocket launched.

Your board is looking at AI deployment timelines measured in months. Your governance doctrine should have been written yesterday. The companies that move on this now will operate at altitudes their competitors cannot reach. The companies that don't will still be running "AI pilots" in 2028 while their rivals run autonomous operations at scale, with full accountability, with doctrine, with moats their late-moving competitors will spend years trying to replicate — and won't.

There is no soft version of this transition. The window to build governance infrastructure ahead of complexity is closing. The checklist cannot be invented on the way down.

Write your flight manual. Before the rocket launches.

The Zero Human Company is not the company with the most AI. It is the company with the best protocol.

Share This Brief

LinkedIn:

In 1935, the best test pilot in the US Army Air Corps crashed the most advanced aircraft on the planet because it was "too complex for one person's memory." His death gave birth to the modern aviation checklist.

Today, four humans are circling the Moon — made possible by decades of obsessively refined flight manuals and human-machine protocols.

The Zero Human Company doesn't mean fewer humans. It means better protocols. The companies writing their AI operating doctrine now are building a moat their competitors won't understand until it's too late.

Full brief: zeroforce.ai/brief/flight-manual-zero-human-companies

X / Twitter (thread opener):

4 humans are circling the Moon right now.

Not because they removed the humans. Because they wrote the flight manual.

That's the Zero Human Company thesis. Thread 🧵

X / Twitter (standalone):

The WHO surgical checklist is 19 items.
It cut surgical deaths by 47%.
The Boeing B-17 checklist saved WWII.
Artemis II is currently 250,000 miles from Earth.

None of this happened by removing humans.
It happened by encoding what they knew.

The Zero Human Company needs a flight manual.

Sources: NASA Artemis II mission updates (April 1–3, 2026); IATA Annual Safety Report 2025; Haynes et al., New England Journal of Medicine, 2009; Gawande, The Checklist Manifesto, 2009; Boeing B-17 history via Wikipedia, Flight Safety Australia, The Aviation Geek Club; WHO Surgical Safety Checklist impact data.