Optical character recognition services

Pick 72312200 when automatic text recognition is the point of the contract, not manual keying. The cleanest test: if a machine reads the characters off a scanned image and outputs text, this is the code; if staff transcribe the data by hand, the parent Data entry services (72312000) fits better.

The boundary that trips contracting authorities is the line between scanning and recognition. Capturing the image is Scanning services (79999100); converting that image into machine-readable characters is OCR. Most real digitisation tenders bundle both, so set the primary CPV code by the dominant deliverable and tag the secondary scope as a supporting code.

The sibling to keep in view is Data preparation services (72312100), which covers cleaning, formatting and structuring data once it exists. Where a contract folds scanning, OCR and downstream structuring into one award, classify by whichever stage carries the most weight in the statement of work rather than splitting hairs across three leaves.

Winners under 72312200 are digitisation bureaux and document-services firms that run high-throughput scan-and-recognise lines, not general software developers. The work rewards production capacity, quality-assurance discipline on recognition accuracy, and a credible chain of custody for sensitive records. If you bid, lead with measured accuracy rates and throughput on comparable back-file conversion, because an authority buying OCR is buying reliable output at volume, not clever code.

Treat this as an opportunistic code rather than a steady pipeline. 28 awards across roughly fifteen years (TED 2009-2025) is thin, and the work clusters in a handful of markets: Sweden, France and Croatia account for most historical activity (TED 2009-2025). Set alerts on the parent Data entry services (72312000) and on digitisation frameworks, since much of the relevant flow surfaces there rather than under this leaf directly.

Much of this work is procured as framework agreements for ongoing digitisation, which favours getting onto the framework first and then winning call-offs. Suppliers active in scanning (79999100) and data preparation (72312100) bid here too, often as one consortium covering the full pipeline.