A proof of concept tests whether a tool can technically do the job. A proof of value tests whether doing it is worth the cost and effort. A pilot tests whether real teams can run it in their normal work. Vendors use the words loosely, so write down which question you are answering before you start.
Three questions, three tests
| Term | Question | Typical scope | Typical length |
|---|---|---|---|
| Proof of concept | Can it detect or do this at all? | One or two repositories or systems, curated test cases | Days to two weeks |
| Proof of value | Is the result worth the effort and cost? | Real code, measured against your current process | Two to four weeks |
| Pilot | Can our teams use it day to day? | A few teams, normal workflow, real tickets | Four to eight weeks |
The lengths are rules of thumb, not standards. Pick what fits your release cycle.
What each one should produce
Proof of concept: a yes or no on capability, with the cases it missed written down. It is the cheapest test, so it is where you drop candidates.
Proof of value: numbers on your own work. For a scanner that might be findings that engineers agreed were real, findings that were fixed, and hours spent on triage compared with before. If you cannot count it, you have a feeling, not a proof.
Pilot: evidence of adoption. Did developers act on the output? Did the tool break a build in a way people accepted? Who asked to turn it off, and why?
Choosing the right one
- Use a proof of concept when you have more than three candidates.
- Use a proof of value when the question is budget or priority.
- Use a pilot when the tool changes how people work, such as a merge gate.
- Combine them for a small purchase, but still write down three separate questions.
Agree the end before the start
Whatever you call it, write the success criteria first. Our post on success criteria before a trial covers how. Also agree what happens at the end: a decision meeting, a conversion path to a paid term if you choose it, and data deletion if you do not.
Do not let a pilot become free production
A pilot that runs three months without a decision is a free production deployment, and it makes the vendor's side harder to plan and yours harder to say no. Set an end date and hold to it.
Try it yourself
Toggle reachability and watch how the finding count changes, then decide what a proof of value should count.
Click through it at your own pace, or open it full screen.
Quick check
Is a free trial the same as a pilot?
No. A trial is access. A pilot has a scope, an owner, success criteria and an end date.
Can the vendor run the pilot for us?
They can help set it up, but your own engineers should judge the output. A vendor run test on the vendor's chosen samples tells you about the samples.
Last reviewed: 19 October 2026.
