# Five Questions Before Your Next Autonomy Decision

Progressive Autonomy Architecture · paa.dev/strategy

Use these before promoting any agentic workflow to less human oversight. Each question has a pass condition. If any question fails, the workflow is not ready to advance.

---

## 1. Can you produce the autonomy record?

For each autonomous action, can the team show when autonomy was granted, what evidence justified it, and which risk it covered?

**Pass condition:** Every promoted workflow has a retained record of the decision, the date, the evidence window, and the accountable owner.

---

## 2. What evidence earns less review?

What signal, over what period, is strong enough to move the workflow toward more autonomy? A good demo is not evidence.

**Pass condition:** The promotion bar is written before the workflow runs at volume, and the team can explain why that bar matches the risk.

---

## 3. Who owns the review discipline?

At cold start, are review decisions consistent, retained, and usable as labels? If not, the maturity curve never gets funded.

**Pass condition:** Review decisions are recorded with enough consistency that they can become training and calibration evidence later.

---

## 4. What sends autonomy backward?

What degradation, incident, or drift condition tightens oversight again, and who gets alerted when that happens?

**Pass condition:** Regression has a written trigger, a safer operating mode, and an owner before the workflow is promoted.

---

## 5. Should this task ever reach full autonomy?

Some workflows should stay partially governed permanently. High-stakes, irreversible, or regulatory actions may warrant durable human oversight regardless of system maturity.

**Pass condition:** The team has made an explicit decision about the workflow's autonomy ceiling, and that ceiling is documented rather than deferred.

---

*PAA is published as a free, vendor-neutral, open standard. Read, use, and adapt without permission. paa.dev*