Skip to main content
Procore

How can I improve the accuracy of data populated in the Document Management tool?

regional availability
The Document Management tool is available in select countries. It is not yet available for Procore accounts in the United States. For more information, please reach out to your Procore point of contact.

Background

To help save time and reduce manual data entry, Procore can recognise and populate data for documents uploaded to the Document Management tool with the use of machine learning, naming standards and project settings. See What are the different fields in the Document Management tool? and What data can Procore automatically populate when uploading files to the Document Management tool?  

In order to reduce the potential for any inaccuracies with automatic data population, the machine learning model was built to prioritise accuracy and will not attempt to "guess" information for the sake of populating information. 

Answer

To help ensure the most information can be populated correctly from files uploaded to the Document Management tool, please follow the best practices below:

  1. Standardise Project Files
  2. Set Up and Follow a Naming Standard
  3. Configure Document Management Fields Before Uploading Files

Standardise Project Files

  • The machine learning technology will look for drawings, specifications and other documents types. We recommend following these guidelines for the best results:
    • PDFs should be in vector format.
    • PDFs should be in horizontal orientation.
    • Fonts used should be standard fonts (simple Sans-Serif/UTF8).
    • Font sizes should be similar throughout the file. 
    • Words should read from left to right or from top to bottom.
    • Disciplines should adhere to one of the following standards:
      • US National CAD Standard.
      • BS EN ISO 19650 Standard.
  • Drawings should have consistent areas for Title, Number and Discipline.
    • The machine learning model will identify a "Region of Interest" (ROI) that it will scan for information such as the title and number. This drawing information typically exists on the bottom right corner, which is where the algorithm will look first. However, the ROI can also be identified based on the contents of the block.
    • Drawing titles and numbers should be labelled on the files as 'Drawing Title' and 'Drawing Number'.
      Note: While there are no restrictions on the length of the drawing title, shorter titles are recommended. 
    • The Discipline is predicted based off the region of the world that the drawing originated from.

Set Up and Follow a Naming Standard

If the filename of a document that you upload to the Document Management tool matches the naming standard set for the project, Procore will automatically populate document metadata based on keywords and identifiers within the original filename uploaded. See Edit the Naming Standard for the Document Management Tool and Automatic Data Entry from the Project's Naming Standard

Configure Document Management Fields Before Uploading Files

Set up Document Management fields and additional Procore fields (such as Project, Location, Originator and Stage) before uploading files so that additional information can be populated automatically. See Manage Configurable Fieldsets and Default Fields for the Document Management Tool and Automatic Data Entry from Project Information.