|Project Cortex Adds AI To Office 365|
|Written by Kay Ewbank|
|Monday, 11 November 2019|
At Ignite 2019, held last week in Orlando, Microsoft announced Project Cortex, a system that uses machine learning to make better use of data held in an organization's Office documents and services.
According to Microsoft,
"Project Cortex applies AI to automatically organize your content, and delivers innovative experiences—topic cards, topic pages and knowledge centers—in Office, Outlook and Microsoft Teams"
Microsoft says Project Cortex builds on intelligence from the Microsoft Graph, a variety of Microsoft AI technologies, and the leading content services of SharePoint. It also builds on Knowledge Network which was an add-on to Microsoft Office SharePoint Server, and relies on all information available to Office 365 being indexed in a central database. What wasn't made clear at Ignite was how Cortex fits with Graph Data Connect, Microsoft's unified programmability model and APIs that can be used to access data within Office 365 and other sources.
The Project Cortex uses AI to gather and analyze content across teams and systems. It can recognize different types of content and extract information from them. That content is automatically organized into shared topics like projects, products, processes and customers. Cortex then creates a knowledge network based on relationships among topics, content, and people. The topic pages can be used to make the information available in wiki-like structures, and topic cards can be used to show the info to users of Outlook, Microsoft Teams, and Office.
Cortex's image and text recognition can identify over 10000 types of attributes and will extract text from images and PDFs. Users can set up forms processing to identify information in forms and extract that information as metadata. Machine teaching can also be used so users can train the AI to recognize information in unstructured documents. Cortex collects content including files, conversations, recorded meetings and video, and categorizes it based on its type, and tags it with extracted metadata.
Cortex then uses AI to 'topic mine' metadata to identify topics and content that relates to them. Topics are actually stored as knowledge entities in Microsoft Graph, along with the relationships between those topics.
An example given by Microsoft was that of someone writing an email. Cortex sees that the email refers to a particular project, so shows you a topic card with other information such as who else is working on the project, what the deadlines are, and any other relevant info.
Cortex will be made available as a private preview, and you can apply now if you're interested.
or email your comment to: firstname.lastname@example.org
|Last Updated ( Monday, 11 November 2019 )|