As part of our 2024 Revenue Cycle Summit, we held a panel about the exciting new ambient speech technologies and the revenue cycle. The following six founders/CEOs of ambient speech vendors that are changing healthcare joined Adam Gale, KLAS’ CEO: Ian Shakil (CSO at Commure), Punit Singh Soni (Founder and CEO at Suki), Rizwan Pasha, M.D. (CMIO at Microsoft), Shivdev Rao (CEO and Co-Founder at Abridge), Nikhil Buduma (Co-Founder & Chief Scientist at Ambience Healthcare), Matthew Ko (President & CEO, Co-Founder at DeepScribe).
This blog gives a glimpse of the discussion and questions answered by these industry leaders.
Points to Know
- Panelists Buduma and Rao propose that revenue cycle leaders, including CFOs, need to be engaged from the earliest stages of selecting ambient technology. That way, the right documentation and metrics are in place from the get-go to show ROI.
- Singh Soni listed the following things to look at when evaluating different ambient speech solutions: the intelligence behind the models, the metrics or ROI that the company has seen, and how they are being adopted. Ko also recommended checking the adoption rate and the change management process of the vendors.
- When it comes to effectively tracking the impact of the tools, Pasha suggests focusing on how you’re making providers more productive and to measure that with surveys and EHR and Signal data.
- Panelists shared their visions of the future. Parts of their visions may seem like they are straight out of science fiction. But similar to how the world changed after electricity and the internet, the introduction of generative AI means these visions may be closer than we think.
Read on for deeper insights shared by our expert panelists.
How should revenue cycle leaders be involved in the selection of ambient technology systems?
Buduma began by flipping the question to the audience. He asked those who were using ambient technology right now whether they also had their revenue cycle teams involved in the testing and selection of the solution. Only a small group raised their hands, and that illustrated Nikhil’s point that while ambient speech technology is intended to reduce pajama time for clinicians, that premise can miss the fact that documentation is key for revenue cycle processes to avoid a big mess later.
When considering the implications for organizations and their teams, Buduma said, “Think about how to engage the CIO and the CMIO, who are generally the buyers of these technologies right now, and help them shape how to think about the impact on the revenue cycle. Make sure that they are part of the RFP. Make sure that’s part of the evaluation framework and the metrics you’re measuring once these technologies are implemented.”
Shakil also advocated for having more revenue cycle leaders in the room at those early meetings. He says, “[These meetings] can often be oversteered by the CMIO and the CIO, and then you’re brought in months later, and it’s really hard to change things. Initially, oftentimes, the metrics are all about one two hours of time savings. . . . But there are so many other metrics that matter a whole lot as well. In particular, metrics around documentation, quality, coding optimization in the hospital setting, . . . [and] denials management. We’re seeing in our deployments [that] those later metrics that I’ve just described are actually the superseding and prevailing metrics that support ongoing expansion.”
Rao agreed that, in the US at least, “We don’t get compensated for the care that we deliver; we get compensated for the care that we document that we deliver. So, whether we like it or not, we’re all revenue cycle companies.” Rao then reiterated that ambient speech technology has the opportunity to take the documentation and coding burden off clinicians’ plates. But it’s also an important opportunity for revenue cycle leaders. He said, “This is the opportunity right now to thread that needle for not just, at an executive level, the CMIO who wants to see the road map . . . [or] the CIO who wants to see this integrated very deeply with the existing technologies back inside the system . . . but [also] the CFO and being able to demonstrate ROI.”
What should these revenue cycle leaders be evaluating as they’re thinking about these solutions?
Singh Soni outlined the following three buckets that leaders should be thinking about when it comes to ambient speech:
- Intelligence. Before a note is even generated, Singh Soni recommends asking about the contextual awareness of the models. He recommends asking whether it considers previous data and work done by other doctors and the context of the patient. It’s important to ask about what is being documented and whether it’s encounter specific.
- Metrics. Where better documentation is in place, people should ask what impact they see in terms of revenue.
- Adoption. Ultimately, everything depends on adoption. It doesn’t matter what algorithms can do if the adoption isn’t there. Singh Soni says, “So, [ask] where are you deployed? Are you deployed across all form factors? Are you deployed across various settings? What specialties do you work on?”
Talking further about adoption, Ko added, “If you follow the typical adoption curve, you can see most deployments settling and flatlining around the 30%–40% adoption in the early majority. But to go further down that stack, what we found to be really important growth levers is the ability to customize the models at the provider level.”
Ko also recommends looking at change management and gaining physician trust. “Aside from looking at the efficacy of the model, you also have to really interrogate what is the change management process of the vendors that you’re looking at. Are they willing to really invest the time to go provider the provider and implement these solutions?”
How can organizations effectively track the impact of these tools?
Pasha recommends that organizations look for a vendor with the staff to help measure and define what is important to them and what they are looking for. He looks for the following three elements that show they’re helping providers be more productive:
- Providers are more efficient. “And what that translates into is probably, hopefully, more patience.”
- Physician happiness. “Is this really impacting their burnout scores? Is this really impacting the way that they’re taking care of their patients and the longevity in your organization?”
- The patient experience. “Are the patients looking at this as an intrusive product, or is this something that’s capturing the information as a better note? All of us can see our notes now . . . so is that note accurate and reflective of what’s happening?”
Once metrics are established, Rizwan says organizations should measure these elements using things like surveys, the EHR, and Signal data. The results speak for themselves: “We’re seeing across multiple organizations now that they’re actually seeing more patients and are less burned out, and the patients feel like the experience was better.”
Where to next from here and now?
After each panelist quickly answered what made their company unique in their ability to solve problems with ambient speech technology, Gale steered them to their visions of the future with this technology.
Ko illustrated that earlier during the summit, there was a slide that showed the many health systems adopting ambient speech today. He believes that in the next 3–5 years, 90% of all patient conversations will be recorded. He advocated that we start looking now at what care looks like in that world. “What kind of data is present in in a world where all these conversations are recorded, and . . . are there ways that we can apply this technology to actually surpass what is humanly capable today?”
Pasha added that in his view the future “is almost like the Star Trek room of the future. You walk into a room, you have smart devices that can recognize you, that can hear you, recognize your voice with biometric signs, and then have a conversation. And the way we envision this is in a trusted, responsible AI environment, where the AI has been trained on responsible AI principles, and also in a secure environment.”
Finally, Singh Soni summed up the moment we’re seeing right now: “Pre-electricity, post-electricity. Pre-internet, post-internet. Pre-AI, post-AI. . . . Post-AI, a whole new set of winners is going to emerge. A whole new set of losers will also be there. The whole world is probably going to be built and rebuilt in terms of a data layer that’s in Knowledge Graph, that’s structured and has a relationship to each other.”
Watch the full panel on YouTube here. This panel was held during KLAS’ 2024 Revenue Cycle Summit. Interested in being in the room for these kinds of discussions at KLAS events? Please learn more on our events page or reach out to events@klasresearch.com.
