There’s certainly been a lot of talk about how AI is going to change healthcare, but, perhaps not surprisingly, talk is still all that most healthcare stakeholders have done.

Writing recently on a company blog, Microsoft’s Chris Sakalovsky, VP of US National Health & Life Sciences, pointed to a recent survey which noted that while “83 percent of health systems indicated they were ‘adopting’ AI,  . . . only 15-20 percent of them could say that they were actually using AI to drive real change in their organization.”

As he sees it, most of what experts and analysts talk about when considering the potential benefits and savings of AI is is “almost exclusively focused on clinical applications of AI – like decision support, diagnostics, chronic condition management, medical imaging, drug discovery, and patient engagement.”

But what doesn’t get discussed nearly as much, he says, are the financial and operational benefits of AI.  Why?  “Because clinical use cases draw more research dollars and more headlines on how AI is changing the practice of medicine as we know it, which draw more conference attendees and readers.  Clinicians, who are often involved in the selection and prioritization of AI use cases, are naturally biased toward clinical use cases rather than financial or operational ones. And, compared to clinical AI use cases, financial and operational AI use cases seem, well, boring and maybe even a bit less noble.”

The problem with that approach, Sakalovsky argues, is that, ironically, that’s the slow way for organizations to begin using AI.  “Compared to clinical AI use cases,” he explains, “financial and operational AI use cases are usually faster, simpler, and easier to implement and cost justify than clinical AI use cases. Financial and operational AI use cases also tend to be easier to implement because they usually don’t require redesigning existing clinician workflows. And financial and operational AI can also be noble when they reduce the cost of care–because that’s one of the best ways to make care more available.”

For Sakalovsky, one easy way to start using AI is by introducing virtual assistants.  “Why? Because virtual assistants can unburden your customers and your staff by freeing up time for them to focus on higher value activities or work at the top of their license.”

A virtual assistant, for example, can help customers navigate a health system or website, while “internal virtual assistant can unburden your IT and clinical staff by freeing up more time to solve problems or care for patients by instantly delivering precise information, avoiding search cycles, or offloading clerical tasks.”

In short, then, if you’re stuck in the AI starting gates or your AI application portfolio is in need of rebalancing, says Sakalovsky, here’s a good place to start: ]Build your own enterprise grade virtual assistant.