In Microsoft’s own words: “SharePoint Syntex uses advanced AI and machine teaching to amplify human expertise, automate content processing, and transform content into knowledge”.
Organisations have a lot of content and a lot of silos of information but aren’t getting as much out of it as they could because it isn’t tagged with metadata or people simple don’t know where to find it. Transforming content into knowledge is the strongest selling point of SharePoint Syntex.
There are two custom AI model types used:
What is SharePoint Syntex Document Understanding?
Document Understanding uses an AI model to automate the classification of files and extraction of information. There are two parts to this model: the Classifier and the Extractor.
The classifier is used to identify and classify documents. Let’s say you want to set up a model that goes through your files and picks out a specific document type, for example, a report. You have to train your classifier to identify those types of documents by showing it what a report looks like. Once you have done that, you can upload any kind of document to your library where the model has been applied and then it will only pick out the reports.
Then you can use the extractor to pull information from these specific document types, for example, the date or financial figures. This means that documents must have text in them that can be identified based on phrases or patterns. Document Understanding is used for structured or semi-structured file formats, for example, Office documents, where we see differences in the layout i.e., letters or contracts, but need similar information to be extracted. All of this is configured in the SharePoint Syntex Content Centre site but can also be set up locally in a library.
We used to need hundreds of files to train an AI model, but one of the advantages of SharePoint Syntex is that you only need to train your model with only 5-10 files, including negative examples. The AI model supports different types of files as well, such as PDFs, HTML files, email files, etc.
What is SharePoint Syntex Form Processing?
The Forms Processing part of Syntex uses an AI builder to create AI models within a single SharePoint document library. It uses machine learning technology to identify and extract key-value pairs and table data. You should use SharePoint Syntex Form Processing for structured and semi-structured file formats, for example, PDFs for forms content such as invoices or purchase orders where the layout and formatting is similar.
SharePoint Syntex – Content Assembly
Syntex was initially launched for classification of data, extraction of data, and compliance. Microsoft has now also introduced Content Assembly – aka Modern Mail Merge.
You can now use an existing document to create a new Modern Template, and in turn, use that template to automatically generate new content using SharePoint lists or user inputs as a data source. This will help you to automatically generate standard repetitive business documents, such as contracts, statements of work, letters of consent, etc.
Where does SharePoint Syntex live within your Content Lifecycle?
Most organisations have a content lifecycle and Syntex can/should be a part of this.
Let’s say you are a bank and person A has just applied for a loan. You gave person A an application form that was completed and sent back to the bank. That would take you to step one of your content lifecycle – content is created and collected through the content creation process. The bank administrator will put this application form into the SharePoint document library and Syntex will pick that application form out and extract all sorts of data such as the amount, the date, etc. The file is now enriched with metadata, and we can add a retention label to say we always need to retain these types of application forms for 5 years. Because of the metadata, we can now also search for these files based on the extracted data.
The loan moves through an approval process and gets approved. We need to send the applicant a contract, so we can use content assembly to create a contract to send out to the applicant to say we have accepted the loan. In turn, that contract will go out to the applicant and return signed, where it can be picked up by Syntex again.
Real-world use cases for SharePoint Syntex
Let’s jump straight into another example. I work for a large infrastructure company, and we have 2 million large A1 drawing plans of infrastructure in our area. Before our engineers can go to any construction site to plan the works, they would have to go to the library to collect the drawing plans, and that could take hours, days, or sometimes even weeks.
Now imagine if we scanned all these 2 million drawing plans and converted them into PDFs to upload them into our SharePoint environment. Then we would approach a Professional Service Provider, like Intelogy, to set up a new Syntex model for us. This model would look at a drawing plan and extract all sorts of information like the location, the asset number, etc. All of this would now be available for search and our engineers would be able to search for the appropriate drawing plans whilst being on site and get it up on their tablets.
Here are another few use cases for Syntex Models:
– Contracts/Agreements – Invoices – Application forms – Statements of Work/Proposals – Forms – timesheets, assessments, etc.
Once you have the metadata extracted from your files, you can use it to create reporting dashboards with Power BI or to build Business Process Automation workflows with Power Automate and Logic Apps. You can explore search solutions through Microsoft Search and PnP Modern Search and even create Topics for Viva Topics.
Until now, a Syntex seat licence was required to work with any document or metadata that had been processed by Syntex, hence why many people weren’t even considering using it.
However, the latest licensing update (January 2022 – Syntex Licencing Updates) for SharePoint Syntex states: “Every Microsoft 365 user will be able to build models to classify documents and extract metadata, access content centres and work with processed documents and their metadata”.
It is still required to have a Syntex licence if you want to process any documents (including uploading content into a library to be processed) or to use any other Syntex features like advanced metadata search, document generation, image tagging, etc.
AI Builder Credits
AI Builder credits are required to train and run Syntex form processing models. Previously it required 300 licenced Syntex users to qualify for 1 million AI builder credits or purchase a dedicated add-on Recently, in April, it was announced by Microsoft that they are now giving removing these requirements and instead every licenced Syntex user will be given 3500 monthly AI Builder credits, pooled at the tenant level up to a maximum of 1 million.
* 1 million credits equal 2000 file pages processed via Syntex Form Processing.
SharePoint Syntex Demo
The demo, presented by Leon Armston, starts at 18:20.
What’s coming to Syntex – Roadmap
A couple things are particularly exciting about the Syntex future roadmap.
Image tagging: you will be able to add images to a library and Syntex will determine what’s on that image and add that information to the metadata of that file and then you can use that metadata within Search.
Content Assembly: At the moment Syntex struggles with complicated layout of tables in Modern Templates due to the formatting. You will soon be able to create placeholders within a table in a Word document.
We will soon be able to add extracted entities to open term set if no match is found.
Intelogy have been closely aligned with SharePoint Syntex since it’s early development (Project Cortex). Intelogy were one of only a handful of preferred Syntex launch partners globally, and are founding members of the Microsoft Content Services Partner Program. Book a call with our SharePoint Consultants to find out more about how we can help you.