Microsoft Build Foretells the Future of Teams
I attended Microsoft Build last month, and bring tales from the future.
For Build, an annual conference for software engineers and Web developers, is where Microsoft promotes its products and the tools, techniques, and directions for its third-party ecosystem that have the greatest likelihood of influencing the future. This is especially relevant in the case of Microsoft Teams, given that Teams, unlike Skype for Business, is first and foremost a collaboration solution platform.
Three focus areas from Build have implications for the future of Teams: artificial intelligence (AI), mixed reality, and Microsoft Graph.
Microsoft seeks to help every developer become an AI developer, thereby democratizing AI. Beyond technology, Microsoft talked about the need for trusted, responsible AI products and practices. To demonstrate this commitment, Microsoft announced a $25 million AI for Accessibility program designed to leverage the power of AI to expand the capabilities of the more than one billion people around the world with disabilities.
As Microsoft is known to do, it has introduced a plethora of AI and machine learning products (all under the new Azure brand). These Microsoft AI offerings fall into three broad categories: cognitive services, custom AI, and AI tools.
Microsoft wants to deliver widely applicable, pre-trained AI services related to vision, speech, knowledge, language, and search. The mechanism to do so is entitled cognitive services. The idea is to provide all developers an easy way to incorporate intelligent algorithms that help applications, bots, and websites see, hear, speak, and understand.
With these "plug and play" AI capabilities, developers will be able to add a slew of cognitive services, including:
- Vision -- facial recognition, object identification, optical character recognition, handwriting recognition, emotion recognition
- Speech -- speech to text, text to speech, speaker identification, speaker verification, real-time translation
- Knowledge -- automated FAQ creation, content personalization
- Language -- sentiment analysis, language detection, text translation, contextual spell checking, content moderation, contextual understanding
- Search -- text, visual, video, news search
With custom AI services, organizations can experiment with and train deep learning AI models on proprietary datasets, and then deploy the developed models to the cloud and edge. With AI, Microsoft has adopted an open approach, providing deep support for its own cognitive toolkit, CNTK, as well as TensorFlow, Chainer, and other neural network frameworks.
Developers can customize and extend almost all of the pre-built cognitive service models to include domain specific identification, translation, and understanding.
Microsoft offers a number of AI tools targeted at both the data scientist and the developer:
- Visual Studio Tools for AI -- an extension to Visual Studio that allows developers to access multiple deep learning frameworks, Azure Machine Learning Studio, and Azure Batch AI, and then build and deploy applications to the cloud and edge using the developed models
- Microsoft Cognitive Toolkit (CNTK) -- as mentioned above, allows the creation of deep learning models using your datasets
- Azure Machine Learning (AML) Studio -- a Web-based, visual, drag-and-drop tool that lets you build and test predictive models without writing code. As part of the AML Studio bundle, AML Workbench assists with data wrangling, AML Experimentation service helps test and refine models, and AML Model Management service assists with version control and deployment of models locally, to the cloud, or to IoT edge devices
- Bot Framework -- a software development kit (SDK) that allows developers to build conversational bots for Teams, Slack, Skype, Facebook Messenger, and more. Additionally, the Azure Bot Service provides an integrated environment to assist with bot development and a serverless bot service that scales with demand
What this means for Teams: Expect Microsoft to infuse AI everywhere within Teams: automatic transcription for meetings (already available), content summaries, meeting attendee identification, sentiment analysis, auto translation (partially available), voice control via Cortana (partially available), automatic action items and reminders, productivity metrics (e.g. MyAnalytics), attendee engagement analysis, meeting health statistics, and greatly improved search. And then expect third parties to leverage the available AI tools to extend the native capabilities of Teams and provide intelligent bots everywhere!
Microsoft uses the term "mixed reality" to describe a spectrum from physical reality to digital reality. This spectrum, as shown below, encompasses augmented reality (AR), provided by devices such as the HoloLens, and completely immersive virtual reality (VR), provided by devices such as the HTC Vive> or Oculus Rift.
When I lead vision workshops, I often define collaboration as "two or more people working toward a shared goal that would likely be out of their reach as individuals." During a Build demo session, using the new Microsoft Remote Assist application and the HoloLens, I was able to repair machinery I had never seen before without any training -- the most direct and immediate example of collaboration I've ever experienced.
Remote Assist allows an expert in another location to see what you're seeing, point, and annotate directly in your field of view (because you're wearing a HoloLens).
A second application, Microsoft Layout, leverages VR and AR to let users design and then validate designs within a specific physical space, such as placement of a new piece of equipment in a manufacturing environment.
What this means for Teams: Teams is and will continue to be the hub and proving ground for mixed reality collaboration, even with SharePoint spaces, announced after Build, also delivering mixed reality content.
Continue to Page 2: Microsoft Graph and the Future of Teams