# Aipom Decision Cycle Redesign

> Redesign a recurring product decision cycle around evidence, context, human judgment, AI assistance, authority, feedback, and measurable learning.

- Skill: `deanpeters/aipom-decision-cycle-redesign` (Agent Skill, multi-file: 4 files)
- Install (CLI): `npx skillmds@latest add deanpeters/aipom-decision-cycle-redesign`
- Raw SKILL.md: https://api.skillmd.com/api/skills/deanpeters/aipom-decision-cycle-redesign/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: Dean Peters (https://skillmd.com/u/deanpeters)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/deanpeters/aipom-decision-cycle-redesign

---


# AIPOM Decision Cycle Redesign

## What Is It

Redesign one recurring product decision from trigger through evidence, deliberation, authority, action, feedback, and revision. Use AI only where it improves decision quality, cycle time, learning, or consistency without obscuring judgment and accountability.

## Why Use It

Adding AI to document production can accelerate noise while leaving the decision cycle unchanged. This workflow integrates context, behavior, evaluation, human judgment, authority, and feedback around the decision itself.

## When to Use It

Use after mapping the current productive motion and establishing foundational controls. Choose a bounded recurring decision with an accountable owner and measurable baseline.

## What It Produces

- Current decision-cycle diagnosis
- Future cycle and human-AI responsibilities
- Evidence, context, evaluation, authority, and fallback design
- Bounded pilot, comparison, measures, and adoption plan
- Playbook and improvement handoff

## Who Should Participate

Include the decision owner, Team Lead, people performing and affected by the work, Product Manager, Product Operations, technical and data partners, and governance partners proportionate to consequence.

## Evidence to Bring

Bring observed cases, baselines, decisions and outcomes, motion maps, context packages, behavior contracts, evaluations, work contracts, exceptions, incidents, and practitioner capacity evidence.

## How to Do It

1. Load supplied context into a ledger and define the decision boundary.
2. Diagnose the current cycle: trigger, evidence, synthesis, deliberation, decision, action, feedback, and revision.
3. Identify the constraint, failure modes, missing perspectives, and hidden labor.
4. Set redesign outcomes and countermeasures.
5. Allocate AI preparation, analysis, recommendation, execution, and monitoring separately.
6. Assign human review, judgment, approval, accountability, escalation, and stop authority.
7. Design authoritative context, provenance, behavior, evaluation, fallback, and incident paths.
8. Run a bounded comparison against the baseline with representative cases.
9. Decide adopt, revise, contain, or stop using evidence and preserved disagreement.
10. Convert the proven cycle into a playbook with ownership and review cadence.

## Facilitation Protocol

Support guided, context-dump, and best-guess modes. Ask first which recurring decision must improve. Reuse prior artifacts; do not re-interview the team about mapped facts. In best-guess mode reduce autonomy and scope, label assumptions, and prioritize evidence collection.

## Decision Logic

1. **Repair the current cycle** when unclear authority, missing context, or avoidable handoffs—not AI capability—cause the failure.
2. **Assist a bounded step** when evidence supports value and consequences remain reviewable.
3. **Redesign the full cycle** when several linked decisions, feedback, and context flows must change together.
4. **Contain or stop** when critical behavior, data, authority, evaluation, or recovery gaps remain.

Prefer the smallest design that changes the decision outcome. Do not automate a decision merely because preparatory tasks can be automated.

## Completion Criteria

Finish with the baseline, redesigned cycle, evidence and assumptions, human and AI responsibilities, authority, controls, pilot results or plan, measures, unresolved questions, adoption decision, owner, and review cadence.

## Key Concepts

- Optimize the decision cycle, not document throughput.
- Faster preparation may increase downstream review.
- Feedback must revise context, behavior, or rules.
- Human accountability remains explicit even when execution is automated.

## Organizational Applications

Use for opportunity selection, research synthesis, roadmap changes, experiment decisions, escalation routing, launch readiness, and portfolio reviews.

## Common Pitfalls

- Starting with a tool feature
- Redesigning the happy path only
- Measuring output speed without decision quality
- Adding review without capacity
- Automating authority through vague approval
- Standardizing before a representative pilot

## Combine With

Use context, evaluation, and accountability skills to deepen weak conditions; hand proven behavior into `aipom-workflow-playbook-builder` and `workflow-to-skill-converter`.

## Assets and Templates

- [Decision-cycle redesign template](template.md)
- [Synthetic worked example](examples/worked-example.md)
- [Weak example](examples/weak-example.md)

## Sources

This workflow is an original AIPOM synthesis of decision design, workflow improvement, human-AI collaboration, and evidence-based product practice.