Microsoft Syntex

Microsoft Syntex enhances the advanced features of SharePoint to help you innovate and automate business processes. With Microsoft Syntex, you can automate the entire content lifecycle.

You can use Microsoft Syntex to create and train models to automate the classification of files and the extraction of metadata from the files. The metadata can then help your organization drive business processes, compliance, and information search and retrieval. In addition, Syntex has an advanced metadata search experience that makes it easier to search in context using metadata in document libraries. With advanced metadata search, you can search on queryable columns (note, only licensed users have access this feature). And with content assembly, you can use Syntex to create templates from your existing documents, point them to a SharePoint List as a data source, and use these to create documents.

A content center is a site template for creating, training, and managing Microsoft Syntex models. The first content center (the default content center) is created as part of the Microsoft Syntex setup. Additional content centers can be created by an admin in the admin center. Models can be organized by groups (such as departments) and can be created locally in teams or comms sites. All models are rolled up to the default content center, where you can view all content centers and models as well as usage analytics for them.

A model is an algorithm “trained” using content and human input to replicate a decision an expert would make with that same information. In the case of Microsoft Syntex, this lets you classify a file of a particular business type and extract specific entity information from it. Microsoft Syntex models are trained to automate these decisions and actions to improve the accuracy and efficiency of your business processes. There are several different model types in Microsoft Syntex including form processing, document understanding, and prebuilt models.

To get current information on Microsoft Syntex licensing, visit our Syntex overview page.

You can also learn more about licensing by visiting the licensing article on Microsoft Docs.

Microsoft Syntex models use content types as the classification object and for defining the schema of entities. When creating a new model, you can create a new content type or map the model to an existing one.

Both model types are generally used for the same purpose, which is to automate the classification of content and extraction of information from them. However, there are differences – and reasons based on file structure and model reuse – that should be considered. For more information, see Difference between document understanding and form processing models.

Pre-built models are another type of Syntex model that enable you to automate the extraction of information. These models are pretrained to recognize documents and the structured information within the documents but instead of creating new custom models from scratch, you can iterate on an existing pretrained model to add fields that fit the needs of your business process.

Document understanding models are trained in SharePoint and stored there. For more information, see Where your Microsoft 365 customer data is stored.

Form processing models trained in AI Builder are deployed in the region that hosts your environment and the datacenter where that environment is hosted. For more information, see Administer AI Builder.

Currently, language support is dependent on the selected model type:

• Document understanding: Document understanding models support all of the Latin-based languages (including English, French, German, Italian, and Spanish).
• Form processing: Form processing supports documents in more than 73 languages. For the complete list, see https://learn.microsoft.com/power-platform-release-plan/2021wave2/ai-builder/form-processing-new-language-support.
• Prebuilt: For prebuilt models, only English language invoices from the United States are currently supported. English sales receipts from Australia, Canada, United States, Great Britain, and India are supported.

Models can be applied to SharePoint document libraries.

There are two different types of models – one for structured content (form processing) and one for unstructured content (document understanding). You apply form processing models for structured content to a library because you start from a library to go into AI Builder to build your model for that library. You apply document understanding models for unstructured content, which you build in the content center, to multiple libraries.

Yes, multiple model types and models can be applied to the same library, with some exceptions, please see https://learn.microsoft.com/microsoft-365/contentunderstanding/prebuilt-overview

To get current information on Microsoft Syntex licensing, visit our Syntex overview page.

You can apply Microsoft Syntex models to Office documents, PDFs, and images. For a full list of file limitations and supported file types, see Document understanding overview.

Explanations are used to help to define the information you want to label and extract in your document understanding models. Microsoft Syntex provides different explanation types to support the different kinds of information you need to extract.

Classifier models are used to automate the identification and classification of a document type. Extractor models are used to identify and extract specific information from them. Extractor models are associated with a parent classifier model.

Either can be trained first, but a classifier is necessary to publish or apply the model. If you plan to train extractors, they can be trained first to be used as explanations for the classifier.

AI Builder credits are required to train and run form processing models. Starting in June 2022 we will provide 3,500 credits per Microsoft Syntex license per month pooled at the tenant level, with a maximum allocation of 1M credits per month.

No, but the allocation based on the number of licenses you have is reset.

Form processing models trained in AI Builder are deployed in the region that hosts your environment and the datacenter where that environment is hosted. For more information, see Administer AI Builder.

JPG, PNG, or PDF format (text or scanned) are supported. Text-embedded PDFs are better because there will not be any errors in character extraction and location.

Trained form processing models are applied to document libraries. Unlike document understanding models, the form processing model is applied directly to the library from where the model was created.

Not currently.

With the Microsoft Syntex add-on, the following managed metadata capabilities are available:

• SKOS-based term set import.
• Push enterprise content types to a hub site, which also adds them to the associated sites and any newly created lists or libraries.
• Term store reports, providing insights into published term sets and their use across your tenant.

Yes, when you create an extractor in your document understanding model, you can take advantage of global term sets in the term store to display preferred terms for data that you extract. For more information, see Leverage term store taxonomy when creating an extractor.

We support three types of data sources in Content Assembly today:

• Input from users: we allow users to add values to “fields” manually.
• SharePoint List: We support associations with columns of a SharePoint Library and document library.
• Taxonomy: We now support using your organization’s taxonomy as data sources for a field.

For more details refer to – Create documents using content assembly in Syntex

List columns allowed for selection during templatization based on column type include:

• Single line of text - Supported
• Multiple lines of text
  • Plain Text - Supported
  • Rich text – Not Supported
• Choice (menu to choose from) - Supported
• Number (1, 1.0, 100) - Supported
• Currency ($, ¥, €) - Supported
• Date and Time - Supported
• Lookup (information already on this site) – Not Supported
• Yes/No (check box) - Supported
• Person or Group - Supported
• Hyperlink or Picture
  • Hyperlink – Supported
  • Picture – Not Supported
• Calculated (calculation based on other columns) – Supported
• Location – Not Supported
• Image – Not Supported
• External Data – Not Supported
• Task Outcome – Not Supported
• Managed Metadata – Supported

You can templatize Text & Tables. Templatizing images, smart Art, bullet list in word documents are not supported.

Content Controls in word documents are also not supported.

Modern templates are stored as Content Types.

.docx and .pdf as output file formats during document generation are both supported.