Futura AI
it
Industry2025–2026Verified project

Automated extraction of dimensions from CAD drawings and ISO conformity checks

A precision mechanical components manufacturer receives technical CAD drawings in inconsistent formats and graphic standards across clients and suppliers. Reading dimensional data, tolerances and checking conformity against the applicable ISO, EN and DIN standards was done manually by the engineering office: slow, repetitive work exposed to transcription errors, with direct impact on quoting and quality.

Analysis

Inventory of incoming drawing types, graphic formats in use and the ISO, EN and DIN standards relevant to each product family. Identification of the critical dimensional fields (dimensions, tolerances, finishes) and of the points where a reading error has the greatest downstream impact, in quoting and in production.

Solution

An automated CAD drawing reading pipeline based on OCR and Vision-Language models, which extracts dimensions and tolerances directly from the technical drawing and checks them against the relevant ISO, EN and DIN standard requirements, flagging deviations and low-confidence cases for review by the engineering office.

Architecture

Vision-Language models for interpreting the technical drawing, OCR for text and tables present on the sheet, a structured knowledge base of applicable ISO, EN and DIN standards for automated comparison, integration with the SAP ERP already in use for quoting and production data, and a review interface for human confirmation of low-confidence cases.

Implementation

Started on a subset of drawings and the most frequent standards, validated accuracy with the engineering office against a sample of real drawings, calibrated the system on the errors found, and progressively extended to other product families.

Measured results

  • 86% accuracy in automatically extracting dimensional data from CAD drawings, measured on the client's real technical drawings
  • Quote preparation time cut from 4-5 person-hours to a few minutes, with automatic regulatory checks and access to centralized company knowledge via the SAP ERP
  • Fewer transcription errors on dimensions during quoting
  • More systematic ISO, EN and DIN conformity checks, with uncertain cases flagged instead of missed in manual review

ROI: Value is measured in person-hours freed from quote preparation — from 4-5 hours down to a few minutes per quote — alongside lower risk of non-conformities caught late in production.

Scope and measurement method

  • Scope: the subset of most frequent technical drawings and the ISO standards applicable to the product families selected for the rollout.
  • Evaluation set: a sample of the client's real technical drawings, with dimensions read and manually verified by the engineering office as the reference for comparison.
  • Metric: dimensional callouts extracted correctly out of the total callouts present in the sample. The denominator also includes the callouts the system did not read, not only those it produced an output for.
  • Low-confidence cases: flagged for engineering-office review and counted as not extracted, not excluded from the calculation.
  • The remaining 14% is why human review of uncertain cases is part of the architecture rather than a fallback.

Client: Project delivered for Gruppo SAG. sagtubi.saggroup.com/en/

Have a similar process?

The same method — analysis, architecture, agreed KPIs — applied to your case. An AI Assessment establishes whether, where and how to act.

Evaluate a similar process