Jobletics grew rapidly across the United States. We had a few rounds of funding and a successful exit acquisition in 2019 from a Fortune 500 company in New York. That was around when the Covid pandemic hit. My wife, who is chief of neurology at Kaiser Permanente in San Francisco, was at that time an attending neurologist at Harvard MGH (Massachusetts General Hospital). We teamed up to start our second company, Reviving Mind, which addresses chronic disease in older adults. As a neurologist, that’s predominantly who she provides care for.

She and some other clinicians came up with a novel care-delivery model that integrated community group support, as well as lifestyle medicine techniques for older adults. It also addresses loneliness and isolation, which is an epidemic for older adults in and of itself.

We also created an associated electronic health record (EHR) to deliver chronic disease management services that was developed to work with the overall care-delivery model and the program. Plus, we created an eight-state telehealth clinic, which I don’t necessarily recommend doing for the faint of heart, but we did it anyway, and that grew rapidly. A lot of consumer patients were signing up for the service. We got into Y Combinator, which is one of the premier investors/venture accelerators in the world, in the winter of 2021, and continued to grow that business. As of last year, we pivoted Reviving Mind more to a SaaS services company, selling into community hospitals and other organizations in rural America, where access to and gaps in care are quite pronounced.

One of the biggest takeaways from that experience for me is that while you can have really sophisticated care-delivery models that clearly show outcome improvements in your patients and modern practice management systems with EHR, we found that continued to cause a lot of bottlenecks in terms of the patient enrolling in the appropriate program and receiving the appropriate care and education they needed. We also found that a lot of that work was being done by front office staff.

Of course, while the front office staff are really looking to engage with patients, we’re asking them to do a multitude of things—to be experts in patient engagement, patient education, and patient success. We’re also asking them to be experts in practice management, system optimization, and optimizing the schedule for all the clinicians. And we’re asking them to be experts in insurance, eligibility checking, pre-authorization, and claims management.

The reality is they simply don’t have the training that practice managers and clinicians have when it comes to communicating these things to patients in an efficient and effective manner, such that the patient thinks, “They have told me everything I need to know about this program. It is going to be a fit for me, and I now understand what my burden is from a payment standpoint, as it pertains to my insurance. They’re also checking up on how things are going, and if I’m adhering to the program.” These are some of the things the patient is expecting your team members to be doing, which is often being done by the front office people instead.

We want to address a lot of these tasks, plus operational and administrative overhead, and other issues we observed in running a practice previously and talking to many practice owner/operators and DSO owners. We hope to solve these problems with Neurality AI.

DSOPro: Describe what this program does to address these issues.

Among all the tasks done by the front office, inbound calling, both during and after office hours, can be high volume. Only a small percentage of calls even get answered. They often go to a voicemail, or to an interactive voice response (IVR) system, call center, or something like that, which obviously creates a poor experience for patients. Our system has a human-like voice AI agent that answers the call and can talk to new patients, triage pre-existing patients, and answer questions, as well as schedule appointments directly into the practice management system. It can also do a lot of the claims management, pre-eligibility checking, and pre-auth related stuff. It can actually call the payer to validate and do all the steps needed to submit an accurate claim, and follow up with all the insurance related steps.

And those are just a few of the tasks the voice AI does. When you think about what we’d want the ideal front office team members to be doing, it also includes proactive engagement with the patient. Most communications, especially outgoing and general communications, is now done through a text messaging service, or SMS. Offices send promotional and marketing related messages to patients, but often there’s no good way to have a conversation with patients who respond, especially if it’s high volume. With conversational AI, customers—practice owners, practice marketing personnel, and DSOs—are now able to create adaptive campaigns, such that the system is actually developing rapport. We’ve developed the AI under the pretense of “patient success.” So, a lot of the characteristics and the persona of the AI we’ve built are based on coaching techniques.

We have taken a lot of the learnings from therapy, where listening and consulting with and educating the patient is critical, versus just hard selling, marketing to them, and getting them to take action. As consumers of healthcare, we don’t want to be sold to or marketed to. It kind of feels icky, right?

Patient’s think, “This is my health; I want to know that you care. I want to know that you’re listening. I want to know that you’re educating me.” This AI does precisely that. It asks how your experience was, and the things the practice, clinicians, and team members can improve upon. Developing rapport and ongoing engagement through that conversation ultimately makes patients feel like they’ve been heard and that they have a real one-to-one connection with that practice. Some of the downstream effects include that patients will schedule follow-up sessions at the appropriate time, they will show up for appointments, and the no-show and cancellation rates will go down dramatically because that engagement and rapport has been developed.

DSOPro: Explain how that is customized. Is that done by the practice using the program or do you set it up with them?

Naturally, not every practice is created equal. They have varying and very nuanced characteristics, whether it’s the type of practice, the clinicians, the services offered, even the type of patient population and where they’re located. So, we created different and novel training techniques for the AI. We can very quickly—in a couple of minutes—ingest a large quantity of the practice information. Believe it or not, 80% to 90% of the information a patient would want to know about a practice, from their services to the clinicians to the cost, etc., is typically on their website.

We scrape the website to ingest that information. Our various computer vision techniques allow us to get very accurate information about a practice, which is all we need to develop rapport, engage, and educate the patient about the why, how, and what of the practice.

The remaining 10% to 20% of the knowledge the AI needs is more workflow related, like scheduling complexity and how to optimize scheduling as well as appointments. The practice will give us that information, and then we train the AI on the different things the practice needs to get done in terms of the workflows. We can have a practice up and running in a matter of minutes, within an hour. It just depends on how quickly the practice managers and the operators can give us customized workflow-related information.

Then, depending on that practice’s needs, there are different workflows and actions that what we call the “super AI agent” will be able to do.

DSOPro: How does the AI agent system improve patient acquisition?

Patient acquisition, retention, and engagement are all based on the patient’s ability to quickly access and learn about the practice, like who the providers are and if it’s the right fit for them. By having 100% available, sophisticated, conversational AI that can pick up calls, or respond to texting in a matter of seconds, the patient will be able to get the information they need and feel confident and comfortable about the services offered. This derisks the conversion rates. The number of patients coming in for services dramatically increases.

Add to that the recall or reactivation use, where you may have thousands of patients, depending on the size of the practice group or DSO, who have not been seen in a few years. The AI can proactively engage those patients, develop rapport, and get them scheduled.

DSOPro: Do people think they are talking to a person, or do they figure it out?

The idea isn’t to fool people into thinking it is a human. We are not trying to deceive them by any means. With a sophisticated ear, you can tell it is clearly not a human. Actually, if you explicitly ask, “Are you a human or an AI bot?” it will respond, “I’m a next-generation sophisticated conversational AI that can help you with anything you need and answer any of your questions to the best of my ability. If there is anything I can’t help you with, I will get a human team member involved so you get what you need.” We’re not in the business of trying to fool people—we’re in the business of giving people a better experience.

More from the Newsletter