What a defect-inspection pilot actually validates before it becomes a production line

Drawing-comparison and defect-inspection work sounds like a single capability, and a short pilot only ever proves the narrower half of it.

What a pilot settles well is whether the platform can perceive the difference at all, on the actual parts.

Comparing a part against a reference drawing or a known-good sample, flagging what does not match, is a yes-or-no technical question a pilot can answer honestly within days: either the platform reliably catches the deviations that matter on this specific part, or it does not.

What it cannot settle is the false-positive and false-negative rate over volume, which only shows up at scale.

A pilot run on a small, curated batch of parts does not reveal how the system behaves on the full variation a production line actually produces, including the rare defect type that never appeared in the trial set. That is a different measurement, taken over a longer run, not a smaller version of the same test.

This sits in the same billing category as this site's other research-project work, priced by the engineering day rather than a fixed scope.

A defect-inspection system is tuned against your own parts and your own failure modes, which is not a known problem with a fixed price until someone has looked at the actual data. 825 EUR per engineering day is the rate, on time actually spent.

A standard configuration still takes 12 to 16 weeks to deliver from a firm order, a separate clock from the pilot itself.

A pilot that proves the inspection works and then discovers the production unit is three months away has proven the right thing on the wrong schedule; running the lead time in parallel with the pilot, not after it, is what keeps a successful trial from becoming an unplanned delay.

A defect-inspection pilot is a real answer to "can it tell the difference," not a smaller version of "will it hold up in production," and the two are different amounts of validation, billed and scheduled differently.