The main Microsoft conference “Ignite” on the topics Azure, Office/Microsoft 365, Office, Yammer, Teams took place in Orlando from November 4th to 8th with more than 25.000 (!) participants.
In a number of lectures and keynotes it was shown which novelties the company will provide for its cloud products and services around Azure and Office 365 from now on or in the near future.
One of the most surprising announcements was the “Cortex” project, Microsoft’s solution for cloud-based enterprise-wide knowledge management based on the Office 365 platform, in which – of course – AI will also be used extensively, presented to a broad audience for the first time.
What is so special about "Cortex"?
According to Microsoft, Project Cortex is intended to “revolutionize” knowledge management within companies, especially in the following areas:
- Automatic retrieval of relevant information
- Automatic indexing, grouping and categorization of detected content
- Clustering and mapping of relations between all contents
- Simplification of the manual correction and adjustment regarding the automatically performed “content refinement”.
What should this look like in practice?
So far not all details of Cortex are known (a productive version is scheduled for mid-2020), also no information has been published by Microsoft, whether this service will be a free component of Office 365 – like teams- or whether to purchase additional modules for use, but what was presented and explained at the Ignite, the following first insights into the functionality of Cortex are possible:
Core functionalities of the "Cortex" project
- Cortex retrieves all existing content of an Office 365 instance of the company and tries to “learn” as much as possible about the content (unsupervised learning) using AI technologies. Then the information is “tagged”, grouped and put in relation to each other.
- Cortex is functionally based on Microsoft Graph and the Office 365 Searchengine. I find this particularly interesting, because Microsoft announced in this context an even more extensive integration of (own) contents outside of Office 365 into the mentioned techniques on the Ignite and thus also own database information from already existing systems could be integrated into the “knowledge power” of Cortex.
- The user should be able to view knowledge information from Cortex in all Office 365 applications (Outlook, Word, Excel, Teams), taking into account his or her access rights, and can call this information directly from the Office application.
- Each user has access to his or her personal Knowledge Center, which can be individually customized and expanded with additional information areas in the form of Web Parts.
- There is is also a central Knowledge Management area, which is automatically created by Cortex in the form of a Wikipage. It will also be possible to adjust the page layout and the displayed content.
- With the help of the tool “AI Builder” from Microsoft’s power platform, the AI capabilities of Cortex can be extended individually for the company without coding, e.g. to improve the extraction of information from certain document types, such as invoices or forms, so that Cortex can better evaluate, categorize and group this data.
To get a first impression of the capabilities of Cortex, I recommend the following recordings of the Ignite 2019:
It is also possible to register at Microsoft to participate in the non-public preview. The registration can be found on the website (along with a lot of other information about the project)
I am very excited to see the first practical tests of Cortex. Should this service really be able to perform the described functions – as presented – also for German-language contents, this would lead to a considerable improvement of the internal knowledge management and would (finally) enable smaller units to access and store their know-how independently of persons.
Is Cortex an option for law firms and legal departments?
Could this technology also be interesting for law firms and legal departments?
What about data security?
At the moment it is not possible to give any reliable assessments. However, if the AI models and the generated metadata, as well as the content stored on the platform, remain within the “own” Office 365 instance and are exclusively hosted in Germany (which should be possible by mid-2020 at the very latest), there is a good chance that Cortex can also be applied in the legal market.
This project is definitely an interesting proposition.
This post is also available in: Deutsch (German)