Microsoft has unveiled a package of cloud software designed for health care systems, starting with a free trial to help the industry weather a viral pandemic that is both increasing the need for technology solutions and putting hospitals in financial peril.
Microsoft Cloud for Healthcare includes tools that allow for digital triage, telemedicine and coordination of care using internet-based services, chat and conferencing apps. The product, unveiled Tuesday as part of a virtual version of the company’s annual Build conference for software developers, is Microsoft’s first industry-specific cloud-computing offering. It’s available now via a preview and free trial for the next six months.
The package combines Microsoft’s Azure cloud service — which helps organizations run programs and store and analyze data – with the Teams communication app and developer tools for creating applications.
Microsoft, the second-biggest cloud infrastructure provider behind Amazon.com, has made the health care industry a focus as it tries to gain customers. Some of the tools Microsoft has been working on for several years, including artificial intelligence and automated chat-bot software for communicating with patients, have gained traction as the Covid-19 outbreak has health-care industry customers looking for ways to track the spread of the virus, predict equipment needs and connect patients and doctors online.
“At this particular time there’s the increased demand to go virtual and have connectivity, while needing to pay attention to the bottom line so they can continue to serve,” said Greg Moore, corporate vice president of Microsoft Health. “This is a time for Microsoft to come in and help what are already our deep partners in health care.” Hospitals have been hit especially hard, with costs rising as they care for Covid-19 patients while revenue plunges because other procedures have been delayed.
Providence St. Joseph Health, a chain with hospitals in seven Western states, has used a version of Microsoft’s health care chat bot to screen patients for Covid-19 and funnel those who need care to a provider or a telemedicine consultation, Moore said. In total more than 1,600 Covid-19 care bots are in use that rely on Microsoft’s software, largely as a tool to reduce strain on emergency hotlines. They’re deployed in 23 countries in cities from Copenhagen to Tel Aviv, including a Covid assessment bot from the U.S. Centers for Disease Control and Prevention.
Meanwhile, St. Luke’s University Health Network and parts of the U.K.’s National Health Service have used Teams chat software for telemedicine appointments. Thirty-four million health-care meetings, including appointments, were held in Teams in the four weeks ended April 25, Microsoft said.
In November, the company announced a new Bookings app in Teams to let medical providers schedule, conduct and transcribe secure virtual appointments through Microsoft’s chat and video-conferencing program. The plan was to try it out as Microsoft usually does for early-stage software – with a couple dozen key customers supervised by the software maker. Then the virus hit and demand accelerated for this kind of product, so Microsoft let some 483 hospitals and providers try the app.
“When Covid happened we weren’t quite ready to make it generally available yet but we knew there was such an imminent need,” said Kristina Behr, a general manager at Microsoft. “We felt a sense of urgency to help.”
Now the app will be released publicly, letting individual doctors’ offices or smaller practices use it. Next, the company will roll out the booking application to other industries. Already financial-services companies are trying it for appointments with advisers, and government agencies are also interested – imagine an online appointment to handle a parking ticket. In the future, Microsoft may blend the app with digital-signature capabilities for such tasks as online mortgage-signing appointments, Behr said.
At the conference, Microsoft also planned to announce other tools for software developers and new products, including:
• A set of three tool-kits for making artificial intelligence models more explainable, fair and private. The tools will be part of Azure’s machine learning service and are also available on GitHub for anyone to use. One kit, called InterpretML, helps data scientists determine why a machine learning model is drawing certain conclusions and which parameters are driving the prediction. That’s key for determining whether an AI algorithm is making decisions based on unfair or discriminatory information. A second tool-kit, called Fairlearn, helps developers understand whether the AI model is relying on data in a biased way. Ernst & Young used it to find that their automated-lending algorithm gave a 15% advantage to males. And WhiteNoise is a new feature for making data more private within AI models.
• Microsoft said it built a powerful AI supercomputer using Azure in collaboration with and for the exclusive use of OpenAI, the research group working on artificial general intelligence that Microsoft invested $1 billion in last year. The supercomputer will be used to train massive AI models.