LAS VEGAS — Last year, generative AI was the talk of the town at the annual HLTH conference in Vegas. It was a breakthrough the health tech industry folks needed — or so they thought — and held the promise of everything from virtually diagnosing patients to conducting surgeries to finding disease cures.

This year, key industry players are adopting a more measured tone. In other words: more practical and less pie-in-the-sky.

Two themes emerged at this year's HLTH, hosted at the Venetian Expo. The first is to help automate mundane tasks to avoid the "crisis level" of burnout. The second is slowing down the response time of AI and its underlying large language models (LLMs).

The latter is a concept being adopted by the tech industry at large, known as "the reasoning phase." As Sequoia Capital said in a recent blog, "It’s not enough for [language] models to simply know things — they need to pause, evaluate and reason through decisions in real time."

Gary Lynch, Verizon's (VZ) global practice leader of healthcare and life sciences, noted the first theme, the need to solve existing problems. (Verizon owns a 10% ownership stake in Yahoo Inc., Yahoo Finance's parent company).

"When you walk the [conference] floor, you see a lot of these new, interesting clinical applications. But we really have to help our hospital partners with the operational efficiencies or we're never going to get to remote surgeries," Lynch said.

Companies such as Verizon and Nvidia (NVDA) are keys to the foundation that the technology needs to be built on, while companies like Google (GOOG, GOOGL) and Microsoft (MSFT) offer a base — like cloud computing and paid access to their AI tools — to create solutions that could revolutionize the industry. They all had a presence at HLTH this year and spoke to Yahoo Finance about the progress made in the industry.

The idea of automating tasks isn't new, but the technology is finally in a place where it is now a reality, experts told Yahoo Finance at the conference.

"Over the last year and a half, a lot of these new platforms ... all these capabilities didn't exist," said Umesh Rustogi, Microsoft health and life sciences data platform general manager.

The most prominent example is AI scribes, which listen to and transcribe patient visits and then identify relevant information, like reasons for the visit and what the doctor advised or prescribed, to add to a patient's electronic health record. (Companies like Microsoft and Amazon (AMZN) have launched these products in the past year, resulting, some experts say, in hours of physician time saved.)