Summarize this content to 2000 words in 6 paragraphs
Sara Vaezy, Providence chief digital officer. (Providence Photo)
“It was a very interesting time,” said Sara Vaezy, recounting the earliest days of COVID-19 and how digital health initiatives at Providence were accelerated by the pandemic.
Five years later, the executive vice president and chief strategy and digital officer at the Renton, Wash.-based healthcare provider is leaning into more interesting times with the rise of artificial intelligence.
On a new episode of the “Shift AI” podcast, Vaezy shares insights into how AI solutions are being implemented in an effort to revolutionize patient care and clinical workflows.
In January 2020, Providence treated the first COVID patient in the U.S., giving the organization an early start in adapting its systems. Tools like DexCare, a multimodal care scheduling platform (now an independent startup), enabled Providence to scale virtual care rapidly as in-person visits declined.
“Virtual care 30x’d overnight. And we were able to ramp up along with it through that platform,” Vaezy said of DexCare, which was spun out of Providence in 2021.
Providence also launched a COVID symptom screener and remote patient monitoring system, which kept many patients at home and reduced emergency room strain.
Now, Vaezy and Providence are building key AI-powered solutions that help reduce clinician workload, improve patient access and communication, and more, while navigating the complexities of privacy and regulation.
And the future of AI in healthcare isn’t slowing down.
“We ain’t seen nothing yet,” Vaezy told “Shift AI” host Boaz Ashkenazy. “2025 is really going to, I think, have a lot of development in agents. They’re getting more sophisticated and more useful very rapidly.”
Listen below, and continue reading for highlights, edited for context and clarity. Subscribe to the Shift AI Podcast and hear more episodes at ShiftAIPodcast.com.
On AI integration across healthcare operations
Vaezy said Providence is heavily investing in AI but has been deliberate about reallocating existing resources rather than increasing cost. Key AI-powered solutions include:
Conversation and Navigation Platform (“Grace”): An AI-powered chatbot that intercepts patient messages, understands their intent, and deflects unnecessary administrative inquiries — reducing patient message volume by 30%.
ARIA (Automatic Real-Time In-Basket Assistant): Helps clinicians manage their inbox messages by using AI to draft responses and reduce response times by 50% (from 48 to 24 hours).
Nuance DAX CoPilot: Uses ambient listening technology to automate clinical documentation, giving doctors one to four hours back daily and reducing after-hours charting.
On responsible AI development
“About 40% of our resources are dedicated to monitoring LLM ops and observability, and 60% are dedicated to feature development. So it’s a lot of resources dedicated to that sort of responsible AI deployment per use case.”
“At some point, data can be identifiable, and we don’t want to risk that. So our answer has been generally no [to sharing patient data with third parties to train AI models].”
On the future of AI in healthcare
“Search is going to be completely disrupted, and I think search within the context of healthcare is going to be disrupted … we’re going to have so much more ability to bring in not only real-world evidence or scientific evidence into decision-making, but also just basic additional context at the point of care.”
On the future of work
“Human beings are more than their work. Hopefully this frees us up to enjoy art and family and society — all of these things in a way that can be hard to do when you’re spending all your time grinding away.”
