The frequently asked questions (FAQ) below have been harvested from our customer and partner interactions in the Project Cortex Preview Program, Office Hour series, and Microsoft Content Services Partner Program activities, as well as internal conversations with our field team members.
Want to learn more about Project Cortex, visit the Project Cortex resource center at https://aka.ms/ProjectCortex. Interested in learning more about SharePoint Syntex, visit the SharePoint Syntex page at https://aka.ms/SharePointSyntex.
What is Project Cortex?
Project Cortex, a Microsoft 365 initiative, that applies advanced artificial intelligence (AI) to empower people with knowledge and expertise in the apps they use every day. Based on customer feedback during private preview, we are delivering these capabilities as a set of unique products and services.
Microsoft SharePoint Syntex, the first product from Project Cortex, uses advanced AI and machine teaching to amplify human expertise, automate content processing, and transform content into knowledge. SharePoint Syntex is available for purchase from October 1, 2020.
When will Project Cortex be available?
SharePoint Syntex, the first product from Project Cortex, is available to purchase for all Microsoft 365 commercial customers, starting October 1, 2020.
In addition to SharePoint Syntex, we'll be releasing other innovations later this year that help you organize information into topics and deliver knowledge across Microsoft 365 apps. We'll have more details to share on Project Cortex topic experiences later in 2020.
Why is Microsoft building Project Cortex?
In a world of rapid change, harnessing knowledge empowers people to act quickly, and organizations to drive efficiencies and cost savings.
Empowering our people to adapt and gain new skills is more critical than ever. Today it takes too long to come up to speed on a new role, a new team, a new project or even a new topic. It’s too hard to find relevant content, answers, and experts. And it’s too hard to connect and manage all your knowledge across organizational boundaries.
With Microsoft 365 we are focused on helping all our customers solve these challenges. With the combination of a massive content repository, vast signals from the Microsoft Graph and advancements in AI, we have a unique opportunity to realize the true promise of knowledge, to empower people, to fuel innovation, to accelerate learning and protect information.
Is there a preview program? If so, what does it entail?
Microsoft opened a private preview program at Ignite 2019 to a limited number of applicants. The preview program is closed and no longer accepting new applicants.
The Project Cortex Preview Program offers select customers access to Project Cortex and engagement with the Microsoft product team bringing the solution to life. By actively testing features and providing constructive feedback over several months, our preview customers are helping us produce a solution that meets our customer needs and reflects potential use cases.
What about pricing and licensing?
SharePoint Syntex is generally available as a user-based add-on for Microsoft 365 plans to Microsoft 365 commercial customers. For more details, visit our Microsoft 365 Enterprise Licensing page.
We’ll have more details to share on Project Cortex topic experiences, including licensing and availability, later this year.
Which users need a license for SharePoint Syntex?
Anyone using, consuming, or otherwise benefiting from SharePoint Syntex capabilities requires a license. This includes the following scenarios:
- Access a Content Center
- Create a document understanding model in a Content Center
- Upload content to a library where a document understanding model is associated (whether in a Content Center or elsewhere)
- Manually execute a document understanding model
- View a library where a document understanding model is associated
- Create a forms processing model via the entry point in a SharePoint library
- Upload content to a library where a forms processing model is associated
- View a library where a forms processing model is associated
To illustrate this with an example. Imagine a recruitment team using a SharePoint Syntex document understanding model to extract data from CVs/resumes. In this scenario, there’ll be multiple roles:
- A model builder
- One or more individuals uploading CVs as they are received
- Recruiters and hiring managers reviewing the files and extracted metadata in a SharePoint library
In this case, all individuals involved in this process will require a SharePoint Syntex license.
What languages will Project Cortex support?
When available on Oct 1, SharePoint Syntex will support English. Project Cortex topic experiences will also support English at its GA. We plan to add support for additional languages next year.
Because Project Cortex uses different AI technologies, including some which are language specific, we must first confirm the AI we develop works in English. After that happens, we’ll test and validate the AI with additional languages.
There are features in Project Cortex topic experiences and SharePoint Syntex that you can take advantage of from day one as we work to support additional languages. For example, you can create your own topics in your language of choice and start building your knowledge network. In addition, other features are language agnostic; for instance, you can perform content processing with AI Builder to help you capture information from forms in other languages.
Will admins be able to turn Project Cortex “On” / “Off” at the tenant level?
Yes, we have a full set of admin controls for the functionality. It will ship “Off” by default so that you are able to plot the right adoption journey for your organization.
Is any data processed by Project Cortex in the tenant held or processed externally geographically to the tenant?
Data in your tenant stays in your tenant. For AI Builder Forms, content is transited to an Azure processing queue while being processed and is purged after processing.
Will Project Cortex be available on government cloud? If so, when is GCC release planned?
Project Cortex products will not be available for government cloud at GA. However, we plan to make them available on government cloud in the future.
In the U.S., we go through a FedRamp certification process that happens twice a year – in April and December. The GA dates of each product will dictate when we start the certification process. The earliest Project Cortex could be available for government cloud is some time in 2021.
Security & compliance
How does Project Cortex protect the security and compliance of content?
Project Cortex is based on the security and compliance of content enforced across Microsoft 365. For example, access, retention labels, data sovereignty, and information barriers are all maintained consistently before and after the activation of Project Cortex products. In addition, organizations can apply even more stringent restrictions on the scope and availability of topic information shared by Project Cortex topic experiences.
Project Cortex and SharePoint
How does Project Cortex relate to Microsoft 365 and to SharePoint?
Project Cortex adds new capabilities to the apps and content you already have across Microsoft 365. It includes abilities to find and share knowledge and expertise in apps like SharePoint, Outlook, and Microsoft Teams, and lets you apply AI to classify and organize content in SharePoint libraries.
How does Project Cortex relate to classic and modern SharePoint experiences?
For Project Cortex topic experiences, topics will be discovered from files and pages stored in modern or classic sites. However, topic highlights will only appear on modern SharePoint pages. These pages may exist in a modern type of site such as a Communication Site or a Group Site connected to a Team. In addition, where a classic site has had modern pages added these will also feature topic highlights.
The creation of models in SharePoint Syntex uses the modern UI. For Forms Processing models using AI Builder, the menu item to create a model is only displayed in the modern library UI. Note that a site based on a classic template can still have modern list and library UI applied. Document Understanding models are only created in a Content Center, which is a modern site.
What are content centers?
Based on the leading content services of SharePoint, Syntex introduces a new experience for managing content at scale, integrating metadata and workflow, and delivering compliance automation – the content center. Content centers supply capabilities to teach the cloud how to read and process documents the same way you would manually. SharePoint Syntex uses those insights to automatically recognize content, extract important information, and apply metadata tags. In addition, you can track the effectiveness of your models with integrated visual analytics.
Learn more about Content Centers and content understanding models by viewing our May 20 Office Hours meeting on content understanding.
How are models related to content types?
The models you create and train use content types to classify content. The information you select to extract from the structured or unstructured files is reflected as columns in that content type’s schema. This is intentional, to provide an “intelligent” content type that can be used to fulfill other use cases, like search filtering, defining file plans to apply information protection labels by schema, etc.
Will the content type gallery replace the content type hub for content type publishing?
No, the content type gallery won’t directly replace the content type hub; there will still be a content type hub. The content type gallery is intended to be a more flexible and more centralized way than the content type hub for you to create and manage the enterprise content types that you want to provide across sites in your organization.
Managed metadata services
How do I get access to the modern Managed Metadata Services (MMS)? Will you be using the current SharePoint Online term store or a new tool? Will there be a migration path?
MMS is part of SharePoint and has been upgraded both on back end and modern UX. It is now a suite-wide (across Microsoft 365) service and is available through Microsoft 365. We have some additional new features that will be part of Project Cortex.
You won't need to migrate your existing term stores; you'll just get new powerful features.
Will these new MMS admin experiences be accessible to non-SharePoint admins?
No, the new experiences won’t be accessible to non-SharePoint admins, because the experiences are housed within the SharePoint admin center. You’ll need to be a SharePoint admin to see the global term store and the content type gallery. We are actively working on modernizing the site collection and site admin experiences. As they are modernized, you’ll see one modern experience throughout all your admin interactions.
With what formats does machine training work?
You can apply machine training to Office documents, PDFs, and images. Our documentation will be clear about the supported file types. While certain Office file types are supported, it might be a more difficult to build models from file types with a lot of layout and formatting nuance. The goal is to support all Office file types, PDFs, and images.
Do you need to create a library for each model?
There are two different types of models – one for structured content and one for unstructured content. You apply form processing models for structured content to a library, because you start from a library to go into AI Builder to build your model for that library. You can apply the models for unstructured content, which you build in the Content Center, to multiple libraries.
How are topics generated and updated?
There are two ways to generate topics: 1) users can create topics manually and 2) AI can detect and generate topics automatically. Once a topic is created, AI will update the topic and related resources as it discovers new information. In addition, users can also manually update topic cards and pages to adjust the topic description or highlight new resources and connections.
Can a company specify a list of seed topics that Project Cortex topic experiences can mine?
Yes, you can create a topic from scratch. You can fill in all the information in a topic page and add it to your knowledge network. The AI behind Project Cortex topic experiences can add to the topic over time. This scenario is part of the human and AI story where machine learning and human curation together can build a better experience for end users.
My end users spend most of day in Microsoft Dynamics 365. Will Project Cortex topic experiences be able to surface topic cards in Dynamics like it does in Office 365 apps?
Later this year we’ll ship a set of APIs which will allow customers to highlight topics and show topic cards in applications other than Office 365. This is an essential ingredient in allowing other Microsoft application teams and our ISV partners to build integrations into Project Cortex.
Can we apply cobranding and custom themes to knowledge center sites?
Because Project Cortex leverages SharePoint, we want you to be able to use all the customization features inside of SharePoint to make this really yours. This is about helping you to populate what you would want in your knowledge center and to be able to use the branding that's appropriate for you and your culture.
What APIs are available and will be available for Microsoft Graph?
Later this year, we intend to ship a rich set of APIs through the Microsoft Graph to let you go beyond Microsoft and beyond search. See what we announced about the developer experiences for Project Cortex at Microsoft BUILD.
Is Project Cortex a new kind of graph, or is it related to the Microsoft Graph?
Microsoft Graph helps supply insights about content and knowledge throughout Microsoft 365, and Project Cortex builds on top of the Graph to create your knowledge network. Project Cortex powers new user experiences for knowledge and content management – including content integration via Microsoft Search, automated classification in the content center, and knowledge sharing via the knowledge center and topic cards, for example.
Microsoft Graph provides a consistent API for addressing Microsoft 365 data and personalized insights and is used throughout the suite in places such as Microsoft Search and OneDrive suggested files.
How is Project Cortex different from Delve?
Delve is also based on the Microsoft Graph. It’s limited to Office documents and is more geared towards connecting people with content than Project Cortex.
On the other hand, Project Cortex automatically organizes, categorizes, and classifies content for you and surfaces it in the applications you are using every day. Project Cortex topic experiences create a visual representation of topics and how they relate to each other, the experts around your organization and the content that is relevant for each topic. SharePoint Syntex also offers machine teaching capabilities to process and organize ingested content – from Office 365 and other repositories, through the Microsoft Search technology and the 100+ connectors to other applications and LOBs.
How is Project Cortex integrated with Microsoft Search?
There's tight integration with Project Cortex and Microsoft Search, and the product teams are well aligned.
Project Cortex brings innovation to the content you have stored inside the Microsoft Cloud. But you also have content that lives in dozens of other repositories around your organization and beyond. That's why we're adding capabilities based on our Microsoft Graph content connectors to integrate remote sources — such as file shares, Azure SQL, or third-party sources such as Box and IBM FileNet — into the Microsoft Graph, making it searchable and usable throughout Microsoft 365.
In the future, we'll couple remote content via our content connectors ever more tightly into the applications and services of Microsoft 365. Today, connectors to more than 130 sources are available from Microsoft, along with our partners. To learn more about Microsoft Graph content connectors, visit https://aka.ms/iwantconnectors.
In addition, Project Cortex and Microsoft Search integration goes further. For example, you can create a custom search results page within an intranet site. You can bring in the data you want and create a tailored content experience for Project Cortex. We see opportunities for organizations, ISVs, and developers to create those tailored experiences bringing in information from Microsoft 365 and third-party sources.
Do we need Project Cortex to use Microsoft Search?
No, Microsoft Search is built into Office 365 and Microsoft 365. It’s not required that you have Project Cortex to get Microsoft Search features.
There will be enhancements to Microsoft Search that will only be available with Project Cortex. One such enhancement would be the ability to see topic cards as part of the search results.
When I enable a new search vertical, how long will it take before the vertical shows in search results? Also, how does security trimming work for verticals?
After you enable the vertical, it might take a while before you can view it. If you don't want to wait after enabling it, you can append cacheClear=true to the URL in SharePoint and Office to view the vertical immediately. For more information, refer to the search results page documentation.
Security trimming is honored by the results shown on the vertical. Once added, the vertical is shown to all users in the given scope, i.e., at the organization or site level.
How do we start using content connectors? Also, is it possible to build our own connectors?
Microsoft Graph content connectors are in public preview and rolling out as part of Targeted release. If you’re part of the Targeted release, you should start seeing connectors soon.
To learn more about content connectors, visit https://aka.ms/iwantconnectors. To learn more about setting up the Targeted release options for users or tenants, see the Set up Targeted release documentation.
Yes, it is possible to build your own connectors. You can use the Microsoft Search indexing API to index data and add custom items to search results in the Microsoft Search experience. You can also use Microsoft Search API to query Office 365 data in your apps.
Check out the Sample CSV connector app to learn how to use the Microsoft Search indexing API to create a connection to the Microsoft Search service and index custom items.
Can the PnP search templates distributed last fall work with these connectors?
We expect the PnP Search WebParts to start supporting the connectors in the near future. Today this content is available through the Microsoft Search API (currently in beta). PnP Search Web Parts will start to use these as well. You can track the PnP Search updates on GitHub.
What are the controls from a company perspective? Who would manage these controls? What audit capability comes with the solution?
We will ship both audit controls and security controls for mined topics and knowledge prior to GA of Project Cortex topic experiences.
How can I ensure that this does not get turned on by Microsoft without my approval?
Customers will need to activate Project Cortex.
If an organization, such as a law firm, has an established DMS, such as iManage, in place, how will this coexist with that data set?
While Project Cortex is going to work most optimally for organizations using Microsoft 365 as their DMS, we understand that topics and knowledge exist outside of Microsoft 365. There are content connectors and APIs available – and more being developed – to help customers integrate the information from third-party systems into the Microsoft Graph.
With connectors and topics, you have the capability not only to bring information into your Microsoft apps through topic cards, but also to make connections between the content in Microsoft 365 and the content coming from external sources – providing your users a more holistic view of the organization’s knowledge.
For more information on Microsoft Graph content connectors, including how to use them, build your own, and sign up for the public preview, visit https://aka.ms/iwantconnectors.
Will we be able to exclude topic discovery based on site classification?
We are looking into excluding sites from indexing based off confidentiality tags. However, we currently don’t have plans to exclude sites from indexing based on private or general classifications. One thing to remember is that the admin can always manually exclude sites. We are looking into ways to make it easier for you to exclude a whole set of sites that match specific criteria.
In general, if you blocked a site from being indexed by Microsoft Search, which excludes it from the Microsoft Graph, then it will also be excluded from topic discovery. You can also further restrict which sites are available for Project Cortex. You can control how much you want to ratchet down the visibility of sites.