Summary of "What is Microsoft Syntex? Watch our demo of Syntex in SharePoint!"

Summary of "What is Microsoft Syntex? Watch our demo of Syntex in SharePoint!"

This webinar provides an in-depth overview and live demonstration of Microsoft Syntex, a Microsoft 365 AI-powered content understanding and document processing tool integrated with SharePoint. Formerly known as SharePoint Syntex, it leverages AI and machine learning to automate document classification, metadata extraction, and document generation, helping organizations manage large volumes of content efficiently.

Key Technological Concepts and Features:

  1. Purpose and Audience
    • Targeted at users overwhelmed by large amounts of documentation who need to automate extraction of key information (e.g., HR contracts, lease agreements).
    • Useful for automating repetitive document processing and metadata tagging.
  2. Core Functionalities Demonstrated:
    • Document Processing ("Understand")
      • Classifies documents into types (e.g., lease agreements, contracts).
      • Extracts relevant metadata such as start/end dates, rental amounts, client names, expiry dates, etc.
      • Supports three types of document models:
        • Structured Documents: Standardized layouts like invoices and receipts, with pre-trained AI models from Microsoft.
        • Semi-Structured Documents: Variable layouts but consistent key info (e.g., lease agreements, NDAs).
        • Unstructured Documents (Teaching Model): More free-form, internally generated documents (e.g., statements of work), requiring training with positive and negative examples.
    • Document Generation (Document Assembly)
      • Automatically generates documents based on extracted data using Syntex’s templating mechanism.
      • Supports both automated generation via Power Automate flows and manual document creation with user-filled fields.
      • Ideal for simple templates like letters or contracts but not suitable for complex tables, dynamic headings, or large data sets (for which third-party tools like Encodian are recommended).
      • Microsoft plans to add features like e-signatures in future updates.
    • Advanced Metadata Search
      • Adds faceted, parameter-based search capabilities within SharePoint document libraries, allowing filtering by metadata fields (e.g., content type, landlord).
      • Enhances retrieval of documents from large, folder-heavy libraries by flattening folder hierarchies and filtering on extracted metadata.
  3. Demonstration Highlights:
    • Uploading documents (PDFs, Word, etc.) into SharePoint libraries for processing.
    • Building and editing Syntex AI models by uploading as few as five example documents for training.
    • Defining extraction fields by highlighting areas in sample documents.
    • Training and testing classifiers to distinguish between document types (e.g., signed vs draft contracts).
    • Viewing extracted metadata in SharePoint columns with confidence scores indicating AI certainty.
    • Using extracted metadata to trigger business processes, such as reminders for expiring certificates via Power Automate.
    • Creating views in SharePoint to flatten folder structures and filter documents by classification.
  4. Management and Governance Features (Briefly Touched On):
    • Importing standard taxonomies (e.g., legal sector terms) into SharePoint’s Managed Metadata Service to standardize metadata tagging.
    • Content type publishing targeted to specific SharePoint hubs for better information architecture management.
    • Reporting on taxonomy usage across content repositories, useful for knowledge managers.
  5. Licensing and Usage Considerations:
    • Syntex licensing is user-based, costing approximately $5 per user per month (or £3.80 GBP).
    • Includes access to templates, premium taxonomy services, and AI model creation.
    • AI Builder credits are required for processing structured documents, with monthly limits (3,100 credits/user, capped at 1 million credits/tenant).
    • Additional credits can be purchased in batches; licensing model and credit usage expected to evolve.
    • Only users who benefit from Syntex need to be licensed, not the entire organization.
  6. Limitations and Future Roadmap:
    • Current document generation is basic, not suitable for complex documents with dynamic content or multiple repeating fields.
    • Search does not yet handle numeric or date filters natively (treats them as text).
    • Microsoft is actively improving the product, including adding e-signature capabilities.

Summary of Tutorial/Guide Elements:

Main Speakers/Sources:

Category ?

Technology

Share this summary

Summarize another video

Video