Agenda Content understanding Automation Demos Q&A - - PowerPoint PPT Presentation

Agenda

• Content understanding
• Automation
• Demos
• Q&A

Joining us today

Sean Squires Naomi Moneypenny Anupam Francis

Content understanding

A process for ingesting digital and image/paper content to recognize, extract and process information.

Project Cortex

Your knowledge network in Microsoft 365

Intelligently categorize content and automate processes Tap into shared topics, resources and experts in the context of your work Connect and manage content to improve security and compliance

Content Services Microsoft Graph & Microsoft AI Taxonomy & Content Types

Project Cortex

Your knowledge network in Microsoft 365

Automatically classify content​ Streamline content processes​ Protect & manage content​ Managed metadata evolved​

Project Cortex architecture

Where end users discover knowledge How end users process new content Microsoft Search Where end users discover and curate connections

Microsoft Graph AI

Where Project Cortex makes connections

Topic pages Topic center Content center Topic cards

Microsoft 365

Project Cortex

Connected content Built in intelligence Managed metadata & computer vision Microsoft 365 Including security and compliance

Where Project Cortex gets data

Capture engines Forms processing Machine teaching Content connectors

Better together: People + AI

New AI services and capabilities to make it easier to build content understanding and classification apps directly into the content management flow of Project Cortex Object detection Form processing Document understanding Manual

Process digital content - photos, scans, receipts, business cards, videos with OCR & text Capture content types and metadata from purchase

• rders, applications, other

structured documents Customize for your org’s forms Taught by your subject matter experts Capture content types and metadata from contracts, resumes, other unstructured documents Smart, right out of the box Data entry and labor- intensive processing Any content

Document understanding

 Using advanced machine

learning techniques train a model to classify files and extract information from them.

 These models are centrally

managed but can be applied to any library across one’s tenant

 Option to queue existing library

items for processing by applied model

Form processing

 Leveraging AI Builder – a new

Power Platform capability – one can create models to classify ingested semi-structured files and extract key value-pair data from them.

 These models are created at the

target library (and not currently available for reuse)

Project Cortex: Strengthening the metadata foundation

Coherent across Microsoft 365 Discover enterprise content based on terms Consistent tagging experience with contextual term suggestions and Auto Tagging Improved enterprise content type discovery, use and enforcement for consistent metadata schemas across tenant Tag content automatically with precise terms for search, workflow, and compliance

New taxonomy & content type user experiences

Modern admin experience for taxonomy & content types Updated controls with improved search, filtering and term suggestions

New taxonomy & content type capabilities

Term store analytics

Reports on term store composition & trends on operations

Content types available from modern library UX; support to push via hub sites

Availability of custom content types; enforcement on new libraries

New taxonomy Graph APIs – coming soon

Preview documentation

Sample code, operations, enumerators and more for Graph APIs

https://resources.techcommunity.microsoft.com/graph-apis-documentation/

Project Cortex Office Hours

Project Cortex Resource Center

Mar 25 Overview Apr 8 FAQ Apr 22 Readiness and Preparation May 6 Knowledge experiences May 20 Content understanding and MMS Jun 3 Admin & configuration Jun 17 TBD We plan to continue Office Hours monthly throughout 2020 Sign up at https://aka.ms/ProjectCortex for details

• n upcoming calls

Visit the Project Cortex Office Hours page to download meeting invites