Deposited application · Version 1.0 · method in development

The HAOP Method

Begin with the work system, not the product. Ten steps from True Function to revalidation.

The HAOP Method: ARECC Applied to AI-Enabled Work Systems · 9 August 2026 · DOI 10.5281/zenodo.21868086

What the method does

Purpose before technology. Work before allocation. Requirements before controls. Confirmation after deployment.

The HAOP Method, Version 1.0 · DOI 10.5281/zenodo.21868086

The HAOP method begins with the work system, not a selected AI product. It establishes True Function, reconstructs work-as-done, decomposes actual functions, locates consequential transitions, and defines verification and grounding requirements before deciding whether any function should be delegated.

A legitimate outcome may be constrained deployment, redesign, retention of the existing arrangement, delay, or no deployment. A method whose every analysis ends in deployment is a justification process, not a design method.

Contribution boundary. ARECC (Anticipate, Recognize, Evaluate, Control, and Confirm) is inherited from industrial hygiene. The deposited HAOP contribution is its translation across human, AI, and organizational performers and its connection to True Function, function allocation, verification, grounding, pause authority, accountability, and continuing revalidation.
Development status. Version 1.0 deposits the application and its chronology. The method itself continues to develop: Revision 6.0.1 Section VIII carries the current statement of the ten steps and their operating return path, and the complete practitioner protocol belongs to the HAOP Workbook, in development.

The ten-step work-design sequence

Establish True Function and unacceptable outcomes

Separate the safety outcome the workflow claims from the metric, dashboard, approval, or representation that may stand in for it.

Reconstruct work-as-done through learning teams

Work with people who have current contact with the work and can contradict the formal representation without penalty.

Decompose the work into performance functions

Specify what is selected, transformed, routed, suppressed, framed, approved, or acted upon along the actual path work travels.

Locate consequential transitions

Define provisional verification and grounding requirements before assigning a reviewer or automated verifier.

Establish design requirements and boundaries

State the conditions, constraints, evidence, permissions, failure responses, and operating limits each allocated function requires.

Develop and compare alternatives

Include redesigned human-only work, constrained deployment, delay, and no deployment rather than treating adoption as predetermined.

Allocate functions among the three performers

Make allocation decisions with the people who will perform the work at the table. Preserve the human throughline where interpretation, adaptation, or consequence requires it.

Select technology against the specification

Choose a product only after the work-system requirements and allocated function boundaries are known.

Engineer and validate controls across all five hazards

Provide Verification Capacity and Anchor Access; preserve friction, slack, constraints, and pause; place gates inside the Verification Window.

Deploy, learn, monitor, and revalidate

Recheck controls, anchors, gate states, staffing, incentives, data, models, permissions, and operating rates as the work changes.

What the method records