A small-scale build or trial designed to test whether an approach is technically and operationally feasible before full commitment. Useful when scoped honestly; dangerous when treated as a soft approval.
A proof of concept is supposed to buy information cheaply. The trouble starts when it buys momentum instead, and the organisation discovers it has committed without ever deciding to.
Last reviewed 3 July 2026 · Free and ungated
Before a pilot becomes a programme, have selected senior operators who have scaled comparable systems examine what the PoC could not show you.
Review this before committingA proof of concept is a deliberately limited implementation, of a technology, a process or a partnership, built to answer a specific feasibility question before budget is committed at scale. It differs from a pilot in ambition: a PoC asks whether something can work at all; a pilot asks how it performs in near-real conditions.
When a scale-up decision is on the table, the PoC result becomes the evidence everyone cites, and that evidence is systematically flattering. The vendor assigns its strongest engineers, the data is cleaned by hand, the users are volunteers who want it to succeed, and the integration shortcuts are invisible in a demo. A PoC that succeeds under those conditions has proven the technology functions. It has proven almost nothing about whether your organisation can run it at scale, with average teams, real data and hostile deadlines. Treating the first result as an answer to the second question is how expensive programmes get approved on thin evidence.
Success criteria written after the PoC finishes, so whatever happened counts as success. No pre-agreed kill condition, which turns the PoC into a one-way gate: money and reputation are now attached, and stopping feels like failure rather than the point of the exercise. And serial PoCs used to defer a decision indefinitely, which is its own form of commitment, to the status quo.