# PoC vs Pilot vs Production

Three lifecycle stages with rising evidence bars: PoC proves it can work, pilot proves it works for real users, production owns it.

## Why the CCAR-P exam tests this

The exam checks whether you match the evidence and controls to the stage instead of treating a PoC's success as production readiness.

## Definition

A **PoC (proof of concept)** answers "is this technically possible?" in a throwaway setting — narrow scope, synthetic or limited data, no durability guarantees. A **pilot** answers "does it work for real users under real constraints?" — bounded blast radius, real workflows, and enough monitoring to learn. **Production** answers "can we own and operate this indefinitely?" — full controls, SLAs, on-call, and an auditable **outcome document**.

The bar rises at each step. A PoC can succeed on thin evidence because its **reversal cost is near zero** — you throw it away. Production demands proof precisely because backing out is expensive. Skipping the pilot is how a demo that impressed a room becomes an incident.

Each stage also carries different **compliance posture** and **stakeholder sign-off** needs. A pilot may run under a limited waiver; production usually cannot. Treating stage boundaries as real gates — each with its own evidence — is the discipline the exam rewards.

## Exam trap

Calling a successful PoC "ready to ship" is the trap; a PoC proves feasibility, not that the controls, parity, and ownership for production exist.

## Commonly confused with

- **canary-vs-blue-green** — Those are how you release into production; PoC/pilot/production are lifecycle stages before and up to that release.
- **migration-parity-eval** — Parity eval is a gate typically required to promote toward production, not a lifecycle stage itself.

## Related

- canary-vs-blue-green
- migration-parity-eval
- reversal-cost
- outcome-document

---

Source: Claude Cert Prep — an independent, unofficial CCAR-P study resource (domain P6). Not affiliated with Anthropic.