We’re now enabling models to be published more broadly – to a site or whole tenant. This makes them available for viewing and selection directly from the document library. You’ll also be able to evaluate several models against your content to aid in selecting the most appropriate one.
A set of explanations with dynamic phrase lists based on the labeled entities or the tokens before/after.
You can now use regular expressions (RegEx) in explanations, allowing model trainers to provide more complex metadata extractor and recognition patterns. This addition includes a set of regex explanation templates that can be used or edited like the other templates.
You can now tag documents automatically with MIP sensitivity labels in document understanding models.
We’re updating our experience to support collections (a named group of documents sharing a similar layout – a new feature in preview allowing one to compose several models into a single one) and extractable table item data.
The Content Center is a new site template that provides the creation and management interface for SharePoint Syntex document understanding models. To increase awareness of these product capabilities and the time to train and evaluate a model’s effectiveness with your own content, we’re enabling this site template to all licensed SharePoint tenants. Users without a Syntex license will be able to build and test models in the Content Center but will not be able to publish models for live usage.
When using Form Processing model, you’ll be able to specify that AI Builder and Dataverse storage can be hosted in custom Power Platform environments. Previously, this was limited to the default tenant environments.
I've been working with Microsoft Technologies over the last ten years, mainly focused on creating collaboration and productivity solutions that drive the adoption of Microsoft Modern Workplace.
Found my content useful?
Buy Me a Beer
HANDS ON SharePoint
HANDS ON Teams
HANDS ON Lists
HANDS ON tek
M365 Admin
