Getting Started with Project Cortex — SharePoint Syntex and Microsoft Viva Topics
This is part 4 in my Top 5 Emerging Trends in the Microsoft 365 Cloud Adoption Journey Blog Series
Overview
Project Cortex has been a hot topic of conversation with many customers and partners since it was first publicly disclosed at Microsoft Ignite in 2019.
Project Cortex has evolved from a conceptual code name into a set of solutions focused on applying Artificial Intelligence to provide knowledge and content services initially focused on content stored in Microsoft Office 365 including SharePoint Online, OneDrive for Business, Teams, and Yammer.
Through the Microsoft Graph and Microsoft Search, including the Microsoft Graph connectors, this will continue to expand to non Microsoft content sources over time.
The first two components of Project Cortex commercially available are SharePoint Syntex and Microsoft Viva Topics.
I have been fortunate to have been working closely with the various Microsoft product teams working on Project Cortex and Microsoft Viva over the past 2 years and have presented on this topic with numerous customers, partners, user groups, and other events including this session recording from October 2020.
This blog post will summarize some of the key examples and practical applications for leveraging SharePoint Syntex and Viva Topics.
SharePoint Syntex
SharePoint Syntex was released in October 2020. Syntex enables organizations to automatically identify, extract, and promote metadata from both structured and unstructured documents and images.
I recently presented a session on SharePoint Syntex at the Microsoft 365 Virtual Marathon event which was recorded and will be published on the event’s YouTube channel soon.
While there are many industries and use cases for this technology, some of the most common ones that my peers and I often work with customers on include:
Horizontal Scenarios Across Multiple Industries
- Purchase Orders
- Contracts
- Invoices
- Accounts Payable Forms
- Human Resources (HR) Benefits Changes
- New Employee Onboarding Forms
Professional Services
- Knowledge Centers
- Expertise
- Statements of Work
Financial Services
- Banking
- Mortgage Loan Applications
Public Sector
- Data Protection and Compliance
Forms and Content Processing
A common use case example is focused around processing purchase orders in a PDF format. This scenario works well when we have some control and consistency over the documents being provided.
This example is setup to process new documents loaded into a SharePoint Online document library. In this case we have PDFs. These documents can be manually uploaded or automatically ingested. For example, we have been working with partners like Fujitsu to leverage their scanner and fax technology to digitize and ingest paper based forms.
In this example, we have leveraged the forms and content processing capabilities in SharePoint Syntex. This enables us to easily integrate the AI Builder and Power Automate capabilities from the Microsoft Power Platform to have a workflow automatically run whenever a new document lands in the library. The workflow processes the document and extracts and promotes the columns that have been mapped and trained in the AI Builder model. For example, the name of the vendor, the requestor, and the total amount.
Document and Content Understanding
Document and content understanding models are great when we are working with unstructured documents such as Microsoft Word documents for things like legal contracts, statements of work, or human resources benefits change documents.
In this example, we have created a SharePoint Online document library to store benefits change notifications that are created in Microsoft Word. We can then leverage a content understanding model to parse the documents as they are uploaded and extract and promote key pieces of information as columns and metadata.
These models are created and managed in the new Content Center site collection template.
We start by defining the model, providing example files for training the model, and determining which SharePoint Online document libraries should be used to process new files for the model.
We then create entity extractors to define what fields the model should extract and provide hints on how to identify the information. In this case, we are looking to extract who the insurance provider is for the benefits change notification.
We provide explanations for what to look for. In this case certain keywords and strings that appear before and after the text we want to extract.
Retention Labels
Both types of SharePoint Syntex models can now be used to automatically apply retention labels to help enforce compliance policies. For example, an organization may want to automatically delete HR benefits change notifications after 2 years.
Microsoft Viva Topics
Microsoft Viva Topics is the 2nd part of Project Cortex and was released in February 2021.
Viva Topics is one of the initial solutions released as part of Microsoft Viva which is part of the Microsoft Employee Experience Platform (EXP).
Viva Topics is focused on connecting people with knowledge and expertise across the organization by automatically identifying and recommending topics and then making suggestions for the topic’s relationship to people, documents, and sites.
The initial focus for Viva Topics is around integration with modern SharePoint pages and Microsoft Search to make it easier to discover new topics. For example, related to products and projects such as “Mark 8” which is a fictitious product for our friends at Contoso.
Later this year, Viva Topics cards will also be available and displayed directly in other Office 365 clients and applications such as Microsoft Teams, Yammer, and Outlook.
Topics and their recommendations are organized in the new Topic Center site template.
A great new training resource for getting started with Microsoft Viva Topics has recently been published and is freely available on the Microsoft Learn site.
Security and Compliance
The next blog post in this series will focus on Microsoft 365 security and compliance. This will build upon my recent posts related to its intersection with Microsoft Teams, Power Platform, and Project Cortex.
