atlas by clearpeople

Project Cortex Roadmap (21 July 2020)

Blog posts Project Cortex Roadmap (21 July 2020)
27 July 2020
View all blog posts
  
Artificial Intelligence
Knowledge management
Microsoft Viva

This is a update to the Nov 2019 blog, and contains latest public domain information related to Project Cortex. 

Introduction

Please note that all the information is provided without warranty or guarantee of accuracy. Microsoft may change dates, features, definitions and priorities at any time - and you should therefore double-check any information provided below before relying on it for any decisions or plans you are about to make. Feel free to contact me to discuss more or to let me know of any details below that are out of date or incorrect.

You may also want to read: Project Cortex: What is it and what is next?

As of July 2020

 

Timeline Release  Feature  Notes by author / ClearPeople
In Preview n/a Microsoft Search: Public previews of search connectors The search connectors are a fundamental part of the process of gathering knowledge from external data silos, such as a CRM system a specialised people database, a CV database or a legacy document management system.
In Preview  n/a Microsoft Search:  Search extensibility with custom search results page, query alterations  Allowing customers and suppliers to create bespoke search results pages relevant to the organisation. For ClearPeople this of course includes the ability to update Atlas search results features in line with the Project Cortex and Search roadmap.
Targetted Release  n/a MMS/Taxonomy: 
Modern Managed Metadata number of Terms boundary increased to 1 million

This is an important increase for org taking taxonomy seriously. This will extend service boundary limits from 200,000 terms across the Term Store to 1,000,000. Read more.
Targetted Release  n/a MMS/Taxonomy: 
Modern MMS admin UX		 This is a very welcome item on the roadmap as the current MMS Admin UX has not had any updates since its introduction with SharePoint 2010. Read more.
TBD  n/a MMS/Taxonomy: Microsoft 365 suite wide experiences for taxonomy management, improved end user controls, term store usage analytics. For any organisations that take tagging (and taxonomy) seriously, this will be a very welcome new feature (and has been a major omission). It will be interesting to follow this roadmap item to understand exactly what is within scope. For example, we would like to see the ability to tag a Team or a Channel against a term in the MMS term store.  
Targetted Release  n/a MMS/Taxonomy: Improved enterprise content type creation & management

 From an IA (Information Architecture) this one has big implications. Again, like with MMS, Content Type management for enterprise scenarios has been a neglected feature ever since the Content Type Hub was introduced. For our Atlas users, it may (hopefully) mean we can incorporate more use of centrally controlled content types. Imagine a Content Type called "Standard Agreement" from which other Agreement content types inherit. We are hoping that Microsoft are maturing the ability to centrally add, update, remove columns and settings, which in this example would mean that if you add a new column (e.g. Agreement Date), it gets pushed out to all libraries using a content type inheriting from Standard Agreements. Effectively this has been possibly for years - but relying on a very temperamental Content Type syndication service. Read more.
TBD  v1 Topic cards & pages: Mining of knowledge, topic card in search, feedback of knowledge system, knowledge focused web parts, compliance with info barriers  Until I know more, I am thinking of the Topic Cards and Topic Pages, as the "visible end result" of the mining process. Manual curation may improve the content/accuracy, but in effect the Topic Cards are intended to appear when hovering over a highlighted topic, and clicking through takes you to the full topic page.
TBD  v1 Content understanding: Content center, including forms processing with AI Builder + custom PowerAutomate template, Machine Teaching, smart ingestion of PDFs and better OCR  The Content Center appears to be like the "catch all" area for a range of the Project Cortex features for the "power user". This is the default go-to location for creating the AI models with the new AI Builder (e.g. to create the intelligent rules for forms processing) and Power Automate. Machine Teaching is really one of the new key capabilities that should get far more headlines. It changes the game when it comes to building models - as you can work off a far smaller set of data, and therefore making it much more relevant for the Knowledge Manager role. Machine Learning on the other had is incredibly powerful, but only when you have far larger data volumes and the capabilities to build data models and training data sets, which in turn requires data scientists and domain experts.
TBD  v1 Microsoft Search: Classic Sites get Microsoft Search For us, and for many organisations that have been using Office 365 for a while, this is a welcome piece of news.The reason is that a lot of work, effort and investment has been put into "classic" SharePoint sites with a lot of valuable content and features residing there. Our digital workspace solution, Atlas, straddles both the modern and the classic experience, to allow organisations to leverage these investments for the foreseeable future. With this announcement from Microsoft it will give some comfort that the modern Microsoft search experience can be extended across the SharePoint Online experience whether in Classic or Modern mode. 
TBD  v1 Microsoft Search: Scoping controls in search box  
TBD  v1 Microsoft Search: Custom query suggestions  
TBD  v1 Microsoft Search: Video vertical (Stream +ODSP)  All I can say here is... "thank you, it's about time". Streams is an incredibly powerful (but underused) Office 365 feature. Part of the reason for it being underused has been the inability to surface the content in Streams up into search results (and search based web parts). The amazing thing here is that Streams already transcribes the audio - meaning you are able to search withing the spoken content of a video.
TBD  v1 Microsoft Search:
Conversations vertical		  
TBD  v1 Microsoft Search: Video & audio transcription including within Microsoft Word  
TBD  v2 Topic cards & pages: 
Further Microsoft 365 integration, more topic types, mapping across metadata types, increase scope of sources including taxonomy and machine teaching		  
TBD  v2 Enhanced business process automation in content center  
TBD  v2 File viewers and annotation, deeper integration with information protection e.g. Auto-classification use of retention label   
TBD  v2 3rd Party Extensibility and APIs This is really an exciting development that has been going on for a little while and which Microsoft has finally revealed more officially. At Ignite 2019 it was really good to see Yusuf Mehdi promote more than 100 3rd parties were announced for search extensions and APIs.  
TBD  v2 Smart document creation & generation  
TBD  v2

Microsoft Search:
Advanced search & extensibility, including Azure Search integration, better use of metadata and attaching workflow to metadata

  

Author bio

Gabriel Karawani

Gabriel Karawani

Gabriel is Co-Founder of ClearPeople, responsible for the overall technical and Atlas product vision. He works closely with colleagues at Microsoft on roadmap alignment and innovative Content AI Services programs such as Microsoft Viva Topics and Microsoft Syntex (previously known as Project Cortex).

View all articles by this author