One of the most exciting announcements made by Microsoft at the Ignite conference this week was the introduction of Project Cortex, the working name for Microsoft’s enterprise knowledge network. As a Microsoft Content Services Charter Partner, Intelogy has been privy to this initiative for a while, but now we can finally reveal our insight. It’s not easy to emphasise just how much of a step-change Project Cortex provides; it really could well be the most significant addition to Office 365 we’ve seen in years.
“As we have thought about people getting work done together, as we have thought about productivity — really broadly defined — for more than 10 years we have had this vision of being able to not only help people transactionally get things done but also allowing them to take a step back and capture what the organization knows and put that to use, put that to work.”
– Jared Spataro, Corporate Vice President of Microsoft 365
What is Project Cortex?
Unlike many workstreams in Office 365, Project Cortex shouldn’t be considered as a separate application. It’s more a series of capabilities which significantly improve the out of the box functionality that is already delivered by the platform. As such, it’s probably best to think about Project Cortex as a knowledge network, an innovation that actively organises your content and derives knowledge from across disparate teams and applications. You will no longer need to go searching for knowledge, instead Cortex will surface it up to you in the context of your everyday use of Office applications on web, mobile and desktop.
An enterprise knowledge network
Topic pages are one of the most tangible examples of Project Cortex capabilities being surfaced. A topic is effectively a collection of related knowledge displayed together in a single location. Different organisations will have different types of topic, with common examples likely to include products, locations, clients, and projects. Imagine for example, an AI driven page which groups together a definition of a project, with key project collateral, and project team members – now imagine that wherever this project is mentioned across Office 365, users will have quick access to a summary of the topic.
Effectively, topics are surfacing your corporate knowledge around the given subject. Behind the scenes, there will be hundreds of derived relationships between all of your topics. These relationships between each of your knowledge entities provide a powerful knowledge network.
So, where do staff actually view a topic? Quite simply, topics seamlessly present knowledge in context in multiple locations across Office 365. Reading an email, page or a document and unsure of the meaning of a corporate term? Simply mouseover the word/phase to view a summary of the meaning in the topic card that is presented. Imagine how beneficial this could be for sharing knowledge internally, especially for onboarding a new joiner.
By default, topics will be automatically identified and created by AI – meaning that you can effectively allow topics to create themselves. However, some organisations might prefer to decide what becomes a topic themselves, providing substantial flexibility to ensure your organisation is able to derive maximum value from your knowledge network. We feel that the best balance can be derived from letting AI do the hard work of scanning content and recommending topics, allowing human subject matter experts to refine and curate these suggestions.
While displayed consistently across different parts of Office 365, topic cards will be managed centrally in a new Knowledge Center – a new central hub for managing knowledge across your tenancy.
It’s worth highlighting the scale and extensibility of Project Cortex – not only is the service capable of processing knowledge from vast amounts of data within Office 365, it also connects to third-party repositories, using the new Microsoft Search connectors. As such, your enterprise knowledge network can extend to content in your File Shares, SQL databases, Salesforce and pretty much any application via API ingestion.
Automated content capture
Using the Content Centers feature provided by Project Cortex, SharePoint sites gain the ability to automatically tag content with metadata. This works by first training an AI model on how to read specific formats of documents and then extracting specific content from files. This extracted information is then applied to metadata columns in SharePoint Online improving the discoverability and organisation of content.
Files with a regular structure can have information consistently extracted from them using rule-based form processing. Imagine uploading an invoice or a purchase order and having key metadata, such as the invoice total, date and PO number automatically extracted and then applied as metadata, without any need for OCR software.
Additionally, Project Cortex can also identify important information in unstructured files (i.e. content with inconsistent structures, styles and formats). By utilising machine teaching underpinned by Microsoft AI capabilities, the context of information can be identified. This exciting ability promises the potential to correctly pick a contract end date from a file containing multiple different dates.
On any other day, this would be a huge announcement in its own right – and has probably been buried beneath the other announcements around Project Cortex, but the Managed Metadata Service is getting a major overhaul. Importantly, the service is being extended from operating exclusively within SharePoint Online, to support tagging of content across Microsoft 365.
Content Types are also being significantly improved, with new ‘enterprise content types’ – these could fundamentally change every SharePoint architecture in the future – allowing global content types and metadata changes. Not only does this finally see the end (we hope) of the Content Type Hub but should also make ongoing content type and metadata changes far simpler for organisations to support as part of business as usual.
This complete overhaul of classification within Office 365 is probably one of the most overlooked, yet most exciting updates we’ve seen announced at Ignite. Something, that will be embraced by the people working across the knowledge and information management community. It’s not without good reason that the announcement of the overhauled Managed Metadata service received probably the loudest applause during Microsoft CVP Jeff Teper’s “The latest innovations in SharePoint, OneDrive, and Office for content collaboration in Microsoft 365” session.
Find out more
Project Cortex will become generally available as a premium Microsoft 365 service in the first half of 2020 (the exact licensing plan has yet to be announced). There is also more information available directly in the “Introducing Project Cortex” blog post written by Seth Patton, Microsoft General Manager of Office 365 Product Marketing.