My Take on Microsoft Ignite 2017 – Part 2: Artificial Intelligence

In this two part series I’m talking about my views on the main technology areas Microsoft has announced to want to focus on in the coming years, which are:

Part 1:  Mixed Reality

Part 2:  Artificial Intelligence


Time for part 2!

Artificial Intelligence, or AI for short, took center stage at Ignite this year. Not only during the vision keynote, but a whopping 120+ break-out/roundtable sessions were either centered around this subject or were centered around how AI was being used in a specific Microsoft solution, like for example the chatbot in Teams. However, I basically could have picked any Microsoft solution as an example, as Microsoft has adopted an ‘infuse AI into everything’-mantra.

But what is AI?

Being a pop culture fanatic, whenever I think about or hear the word AI, Terminator 2 comes to mind, one of the movie gems of the early nineties. The movie is about a humaniod cyborg, who has come back from the future to protect a ten year old boy from a more advanced humanoid cyborg sent by the machines to kill him. The boy is destined to lead humanity in a war against the machines in the future. A future in which the AI infused machines have become sentient, achieving true intelligence and have taken over the world, trying to eradicate humanity.

What made this movie and its predecessor so great is that it played right into our fear of becoming obsolete/ losing our purpose….And as people we can’t live without having purpose. It’s the reason we get out of bed every morning.

Let’s look at how AI is being described:

Artificial intelligence (AI, also machine intelligenceMI) is Intelligence displayed by machines, in contrast with the natural intelligence (NI) displayed by humans and other animals. In computer science AI research is defined as the study of “intelligent agents“: any device that perceives its environment and takes actions that maximize its chance of success at some goal.[1] Colloquially, the term “artificial intelligence” is applied when a machine mimics “cognitive” functions that humans associate with other human minds, such as “learning” and “problem solving”.[2]

source: Wikipedia

During the vision keynote and multiple break-out sessions I attended, it was mentioned that Microsoft has spent the past 25 years working on AI, but always in separate product teams. So a team for Bing, a team for Cortana, one for Robotics and so on and so on. A year ago they decided to gather all that scattered intelligence centered around the same subject and form a dedicated research group, led by Harry Shum: which was given the name:

Artificial Intelligence Research Group Microsoft.


To prevent ‘the machines from taking over’, Microsoft has defined design principles for their AI offerings, influenced by ethics, which can be explained as:


Humans are the heroes – People first, technology second.

Balance EQ and IQ –  bridge between emotional and cognitive intelligence.

Honor societal values – Design to respect differences and celebrate a diversity of experiences.

Know the context – no further explanation necessary.

Evolve over time – Design for adaptation.

The first taste we got of Microsoft wanting to ‘infuse AI into everything’, was that during all the sessions everything that was being said by the speaker was translated live and also shown as closed caption, so in real-time. Even more impressive was that it was being translated into 12 different languages using Microsoft Translator technology, a product of the recently formed AI and research team. So in all the sessions, during all the keynotes.

Check out this YouTube video to see the Microsoft Translator solution being used in a classroom:

Microsoft Translator in the Classroom

What Microsoft and other tech giants like Google with their Pixel buds are doing is to try and take away one of the barriers that’s been traditionally hard to break: language.

And it didn’t end there: an other example worth mentioning was how they integrated information gathered from LinkedIN into Office 365 services, such as SharePoint and OneDrive for Business, making it possible to get insights about people you work with both inside and outside your organization using AI, directly from within the opened solution

Awesome stuff, but how do I get started?

Get a subscription to Microsoft’s public cloud – Azure! This will give you access to the Cortana Intelligence Suite, which is a group of components and services with a high level of interoperability and integratability, with a heavy emphasize on gaining insights from (huge amounts) of data.

Here’s an overview of what’s currently available within this suite:


source: SQL Chick

The services that are currently available on Azure are the Cognitive services, the bot framework and Machine learning (ML), which can be integrated into exisiting Microsoft solutions using the several different Microsoft API’s (or Graphs) available, or by creating new intelligent apps from scratch on Azure, either via the App Service or using Logic Apps.

During one of the breakout sessions the integration and ease of use of the Azure Machine Learning components was shown. An Azure Machine Learning dataset was used as a data source to gain insights about certain geographical information using Excel to present the analyzed data. Without having to install anything (at least that’s what was said, haven’t found the time yet to verify it) Excel had AZUREML functions readily available. 2017-09-25 14.51.46

On the infrastructure side the most notable mentions where the new version of SQL Server now with ML capabilities built-in and the multi-geo database solution Azure Cosmos DB – with which you can distribute data across multiple Azure regions, made possible by scaling and replication, promising low latency across regions.

We need more power…

AIP - platform

Analyzing and visualizing huge amounts of data and achieving (and maintaining) high performance and processing rates takes a lot of compute power. With the ‘new-ish’ N-series Azure VM templates, powered by NVIDIA Quadro GPUs, you’ll be able to fulfill your high performance compute need and even train and deploy custom AI models in parallel with the new Azure Batch service. This service uses clusters of GPUs to run experiments in parallel and at scale to reduce training time.

The demand for intelligent (business) apps is on the rise. With the improved API’s, the fact that you’re able to leverage multiple, both internal and external, data sources and integrate that into existing Microsoft solutions or new ones is very powerful. Because of this power and the ease of use, I expect that we will be creating more and more solutions that either are:

  • able to consume from existing Microsoft solutions
  • have a Microsoft solution as a building block to create intelligent apps that add business value

Not having to start from scratch for everything you build, is in my opinion a good thing. Having good API’s and new frameworks adds to that.

Mark my words, your best friend will be a ‘bot’ in the near future. And now you have the components and compute power to build your own!



One thought on “My Take on Microsoft Ignite 2017 – Part 2: Artificial Intelligence

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s