Automated SharePoint Metadata Extraction for Effective Governace.

SharePoint Online and increasingly Microsoft Teams are used for teamwork and collaboration. In my experience, a significant portion of the business information created/received as part of a business activity lives in un / semi-structured documents and other file formats. In many of these situations, we can significantly improve decision making by extracting important information and applying a structure for automation and further processing.

For example, in the below video blog I have picked up an example for a leading oil and gas company, the lubricants team performs daily tests on lubricant samples and the results from multiple vendors are returned in different files formats. In most cases, the team is mainly interested in the Sample code, and the result summary to then match the code and automate the updates to their structured data system where the rest results are extracted and used for further engineering activities and decisions.

In the above use case, automated entity extraction can significantly improve the process and the timeliness with which results are processed to support further business process activities.

I am using SharePoint Syntex to create a 'Document understanding' model to extract key entities and let the SharePoint Syntex model perform the initial processing and populate the SharePoint metadata.

The key steps involved in this process are as follows:

1. Create and Train the Model centrally working with a SharePoint automation expert and a business user with expertise in interpreting the sample data

2. Create and Train the model to extract the key information in specific SharePoint site columns

3. Test and tune the model with your business experts

4. Publish the Model to target SharePoint sites/libraries

Once the SharePoint Syntex model is published work related to extracting the metadata happens in the background, our users can focus on working with the information and leveraging the extracted information for further processing.

SharePoint Syntex models are a great way of democratizing often complex machine automation and build trust with your users. I will be working on my next user case and put SharePoint Syntex to test for extracting information from emails stored in SharePoint Online. This would be a great way to enable an out of the box integration between SharePoint Online and Microsoft Exchange Outlook, users can drag and drop emails directly from Outlook to SharePoint, and Syntex will extract key metadata including email headers into SharePoint site columns.

You can contact me at vivek.b@infotechtion.com for any questions, and especially if you have scenarios in which you are wondering about the automation possibilities in SharePoint Online with Syntex.