Microsoft
Eric Boyd, corporate vice president of Microsoft's artificial intelligence platform.
- Microsoft on Monday introduced new tools to make artificial intelligence easier for companies to use.
- The tools are improvements to the Microsoft Azure cloud's AI platform intended to make machine learning more accessible for developers at all skill levels.
- Microsoft Azure AI chief Eric Boyd said the company sets itself apart by thinking about "how to make this stuff really simple."
- Simplifying artificial intelligence and machine learning for cloud customers could make Microsoft more competitive in the fierce cloud computing battle with market-leading Amazon Web Services.
- Artificial intelligence is one of the technologies Microsoft CEO Satya Nadella recently said will play a key role in the company's future.
- Click here for more BI Prime stories
Microsoft Azure AI chief Eric Boyd said the guiding principle that sets the Redmond-based company apart when it comes to artificial intelligence is "how to make this stuff really simple."
With that in mind, Microsoft on Monday introduced new tools to make artificial intelligence easier for companies to use, including making the machine learning features of its Azure cloud platform more accessible to developers and users of all skill levels.
Artificial intelligence is one of the technologies Microsoft CEO Satya Nadella recently said will play a key role in the company's future. Microsoft's AI platform, Azure AI, is at the center of it. The platform now has 20,000 customers, and more than 85% of Fortune 100 companies have used Azure AI in the past 12 months, the company says.
A big opportunity in keeping it simple
Simplifying artificial intelligence and machine learning for cloud customers could make Microsoft more competitive in the fierce cloud computing battle with market-leading Amazon Web Services.
Azure AI started by taking Microsoft tools it has built for use internally, and makes them available to customers.
The idea behind Azure AI to create tools that help companies accelerate the work they do, and make it easier to do that work with a limited number of data scientists or specialists. Rather than start from scratch, developers can take what Microsoft has already accomplished, and do the relatively minor amount of work to customize it for their needs.
"If you want to do speech-to-text, you could hire a bunch of data scientists and train your own speech-to-text model, but it will be much better to just use the service we already have," Boyd said.
The business is a "major investment" for Microsoft, said Boyd. He declined to reveal the number of employees working in Azure AI, but said it's "thousands of people across the company."
A new user experience
The set of tools announced on Monday includes what Boyd characterizes as a new user experience that makes it easy for people at all skill levels to build machine learning models - the complex math equations that underlie the algorithms that power what we know as artificial intelligence.
Boyd was referring to Microsoft's redesigned ML Designer, a tool that lets developers use a simple graphical interface to put together models using existing data sources, with little to no code required.
Beyond the ML enhancements, Microsoft also added other tools to make it easier for customers to use AI in their own wares. For instance, it's launched a Bot Framework Composer, which lets customers use a graphical interface to code their own chatbots and automated customer service agents.
Other new announcements include Personalizer, an AI-powered tool to help its customers build systems that recommend content or webpages to individual users - kind of like YouTube's vaunted algorithm, or Amazon's product recommendation features.
Microsoft also introduced a new update to its Text Analytics service that can detect personally identifiable information (PII) from documents, which could help hospitals or other regulated industries build software that automatically blanks out sensitive information.
The bigger picture
Microsoft has an advantage when it comes to AI, Boyd said, because it knows exactly what enterprises need from working with so many thousands of them over the years. That's translated into helping customers take advantage of cutting-edge AI in their own lines of business, he said.
Energy management company Schneider Electric, for example, uses Azure machine learning to train models that have shorten the time it takes company scientists to analyze data they collect from pumps all over the world from one month to one day, Boyd said.
Lexmark, a printer manufacturer, uses Azure machine learning to build models using data it collects from printers to figure out when the printers should signal an ink shortage to minimize downtime.
Amazon Web Services, Microsoft's chief rival, started out mostly providing services to startups, though has grown into a fierce competitor for larger customers. Microsoft, meanwhile, sees a real differentiator in this deep experience in working with even the most established companies.
"We have learned what are they things that they need and what really works for them," Boyd said.