Episode 112: Using AI Voice to Support Nurses and Improve Care with RJ Burnham | EP112

Right now there's no good way to automate that workflow. This is a key workflow that exists across everything and is all done by hand. The problem we're trying to solve is to make this a reproducible engine that can drop into the healthcare systems as they exist today because the last thing you want to do is give nurses and doctors another tool to administer and work through since they're already overloaded with technology. So we're really trying to embed it in the existing platforms that they already use today and integrate it that way.

Hi, my name is Demetri Panici and I'm a content creator, agency owner, and AI enthusiast. You're listening to the AI Agents podcast brought to you by Jotform and featuring our very own CEO and Founder, Aytekin Tank. This is the show where artificial intelligence meets innovation, productivity, and the tools shaping the future of work. Enjoy the show.

Hello everyone and welcome back to another episode of the AI Agents podcast. In this episode we have R.J. Bernham, the founder and CEO of Conig. How are you doing today, R.J.?

Doing well, Demetri. Thank you so much for having me on today. Yeah, excited to chat. First of all, let's just kick things off by getting a little bit of insight as to your background and how you got into AI in the first place.

Oh, so I have my like I mentioned, my name is R.J. I was founder here at Conig and I've actually spent my whole career building communication platforms. It's been almost three decades working in the field of what we now call voice AI. In the past it was really called telecom or speech recognition or natural language around it. It's been kind of fun to spend the career pushing the limits of what's possible with voice.

Very cool. And you obviously guys are in the healthcare sector, right?

Yeah, we've really kind of focused in around the problems in healthcare. I really look at it as I've spent my career building more platform type companies. When we started Conig, we started out with the mission to understand problems in a couple industries and time and time again we kept coming back to healthcare. It really comes down to I feel like healthcare is one of those places where communication just sucks, especially as a patient or care person. I don't come from that industry or world, but it's been interesting to break into this.

My core piece is I've been a pro-level healthcare user. I have a daughter with special needs and I've done a lot throughout the healthcare system. I've learned to be an advocate and how to fight, follow up, and chase things down. But it's way harder than it should be. As I started looking at it with the background of voice, I came to the conclusion that maybe this is an industry where we can actually make it better for both sides.

One of my frustrations with voice AI, IVR, and communications technology is it's usually heavily focused on just doing cost savings for the company deploying it. You always look at how to just automate conversations. Healthcare is one of the places where I actually think it can do a better outcome both as a patient and from a technology and communication perspective for providers, payers, insurance companies, and hospital systems.

We have a problem in healthcare: there's not enough nurses, not enough providers, and a huge labor shortage in this field. So much time right now is wasted on the stupidest things. That's what drove me to say, why don't we tackle this and take it on?

Yeah, that makes a lot of sense. So what do you think is the way that you guys are specifically handling it and helping the people who are dealing with these problems? Maybe if possible, a scenario would help. I remember you went through an example scenario on our pre-screen and it was pretty good. I'd love to hear a little bit more about how you guys are helping and maybe the audience could benefit from something like that.

The thing that's happened over the last piece is I spent the first stage of the company basically doing a ton of problem discovery. I connected with people on LinkedIn and just asked them to talk about the problems they see in their business and scaling it. I didn't want to sell anything, just wanted to learn. I ended up having probably 50 to 60 of these across healthcare and we started to see a common pattern: a lot of time is wasted actually getting a patient on the phone to have a conversation.

Maybe this is a post-discharge follow-up after someone's been in the hospital. Nurses usually have a big stack of paper cases they're supposed to check in with. How often does anyone ever get a hold of you in those cases? Almost never. It doesn't happen. Maybe it's making sure that after your physical you actually did the exams and pieces around it. Maybe it's an appointment reminder. No matter what you call this, whether it's in home healthcare or telehealth, they all have this same core reproducible workflow around getting a patient on the phone and actually getting them to connect.

I don't know about you, but I don't answer most of the phone calls that come into my phone anymore if I don't recognize the number. That's the very first challenge. Add to it, we have things like call screeners that make it really hard for automation to make their way through. In healthcare, we have to make sure the person we're calling is really the right person. Am I really talking to Demetri? Can I validate stuff before sharing health information? There are a lot of rules and regulations to protect that for good reason.

Lastly, if I got through all those gates, can you actually talk right now, Demetri? Are you in the car? Are you distracted? Is there a better time? All of that can easily take three to four minutes to get through. Right now in healthcare, they have people doing that. Nurses, the most resource-constrained person in healthcare, are having to do that workflow. It's crazy. It blows my mind. I come from automating and scaling outbound calls and go back and say, why are you doing this?

We had a great workshop working with Blue Cross Blue Shield of Florida. They actually know with high certainty who is probably going to be readmitted into the hospital after an ER visit. They know from data that if you've been to the hospital twice in about two weeks, there's a 50% chance you'll go a third time. If you go there, there's a 50% chance you'll go a fourth. It actually goes all the way down to five or six. They have 100 nurses on staff doing outreach and trying to intervene around this.

The problem is when they did the analysis, they saw 60% of the time they spend is doing the wasted conversations I outlined. This is an area where we really want to help automate portions of this because it's absolutely solvable. Let's have nurses, instead of having 15 clinical conversations, have 30 and really improve outcomes. That's 30 patients each nurse can then follow up with, which before would have been half of that.

I think a lot of people tend to have an aha moment when you talk about these types of things. Personally, when I discuss it and as I've done this podcast more, every person I talk to about healthcare issues, I think it's funny to think about all the conversations I've had complaining about healthcare and the context behind what could be better. I feel like it's always not actually in the conversation. We're complaining but we don't know what could be better or what issues are causing it.

As I've interviewed more people like yourself and others in this space trying to help healthcare, it seems apparent they're struggling for time and a lot of it is nonsense paperwork. You're probably providing one of the first communication solutions. I haven't had that yet. I've had other solutions more about helping with paperwork and that kind of thing, but this is definitely the first I've seen with communication, which I think is awesome.

So what is your thought on how you can do this at a practical level? What has the reception been like for clients you've helped and how did you get a footing in the market in general?

The first thing is trying to describe the problem the right way and the incentives within healthcare. When I first started working with healthcare, the challenge is they speak a foreign language. Coming in as a non-healthcare person, there's so much terminology to understand. There's also a dynamic balancing opposing incentives depending on whether you're selling to providers, doctors, and nurses or to insurance and payers.

What you really have to do is figure out the real impact and how they are compensated, what the financial model is underneath. That's tough as an outsider. That's been my biggest challenge breaking into healthcare. I've been lucky to get some great advisers who helped navigate that. The first thing you need to do is describe to them why this matters.

If you just tell them you're going to save nurse time, they get it but it doesn't click. If you connect it to things like star scores that explain reimbursement rates for Medicaid or Medicare, that pushes it in. Healthcare is very much a business, especially in the US.

I've had the opportunity to see healthcare in three different countries: the US, Portugal, and Brazil. There are pros and cons to all. The US system is a financially driven machine. I don't know how to put it, but I wouldn't call it efficient. I've had funny experiences where a disgruntled receptionist doesn't seem to want to help me, which is disconcerting when it's about my health.

When I talk to people, it seems pretty universal. Maybe they're overwhelmed with too many things and don't know all the details. The only thing they do well is ask for your money accurately. Family members who've had surgeries haven't been asked about follow-up appointments, which seems psychotic. Many places don't have automated sequences to make patients feel cared for.

If healthcare is a business wanting to make money, it's worse they don't do nurturing things that standard business practices do on recurring sales bases. You make sense?

Absolutely. There are changes driving portions of this behind the scenes. I think we'll see improvements as things transition to more outcome-based payments for providers and insurance. It's tough from a reimbursement perspective, but the core use case is when you go to a doctor and have your follow-up, nobody ever follows up. When you're sick, you're at the least capable stage to be responsible and think about follow-up. Much gets dropped because you're not operating at 100%.

If someone actually followed up and had a conversation saying, 'Hey, how was it going?' not just reminders, that would help. We all get texts and emails and ignore them because we're overloaded with notifications. The core portion is to actually have a conversation. That's what we're trying to build. I want to scale more healthcare conversations, not just automate things.

I think the conversation would make me feel better if I got a phone call and then an automated email. The problem is right now there's no good way to automate that workflow. It's a key workflow that exists across everything but is all done by hand. That's the first problem we're trying to solve: to make this a reproducible engine that can drop into healthcare systems as they exist today.

The last thing you want is to give nurses and doctors another tool to administer and work through since they're already overloaded with technology. So we're embedding it in existing platforms they already use and integrating it that way.

So far it's been really good. We've looked at how we partner with providers. We're focusing on specific subsegments. Dental is a great place. It's probably one of the areas most willing to jump on stuff. Dental is a recurring revenue business because you have to go back. If you miss an appointment, you quickly become churn. If you cancel, the odds of rebooking go down tremendously until you have a problem and come back.

I've noticed dentists book your next appointment before you leave the chair. That's operationally to get you back in and keep you in the system. Dental has seen the most adoption of technology and is thought of almost like software subscriptions.

Just to follow up on dental, it's kind of a required recurring thing. Technically, no one has to go to the dentist, but they get you on a recurring model by scheduling your next appointment. In some cases, like major surgery, patients weren't asked when to schedule follow-up, which sounds psychotic.

Many places especially outside dental don't have routine follow-up. Dental is relatively routine and low risk. Major surgeries and ER visits often don't have case workers assigned or they're overloaded and can't get a hold of everyone. That's where it usually gets dropped and what I'd love to change.

Could you talk more about your experiences and how it's going right now?

The biggest portion we're exploring with Conig is partnering with places focused on patient experience and communication. We're leveraging what we've done with the plug-and-play program to do problem discovery. In payer environments, which means insurance companies, the providers are doctors and nurses. The first thing is to find a good champion who understands the problems so we can partner with them to build a model that automates part of the workflow and changes the number of clinical conversations from X to Y.

This also reduces provider burnout because nurses didn't sign up to leave voicemails. They want to talk to patients and improve lives. Taking that off makes them happier. Our first customer, SQM, was in charge of patient follow-up for insurance companies. They do about a million patient follow-ups a year by hand. We automated the first portion of the outreach flow and saved them the equivalent of 10 full-time employees, saving over $700,000 a year.

This isn't low-wage call center workers; these are healthcare professionals. Taking off the worst part of the job improved employee retention. The data showed 80% of the time when they connected, it wasn't a good time and they needed to schedule a callback. People often say they don't want to talk right now. We automated 80% of the calls they were doing by hand.

That's pretty cool. Are these problems more universal than we imagine? I know we mainly talked about the US. Do you have any insights from Lisbon or other countries?

I believe it's universal but dynamics differ. Different countries have different systems, metrics, and behaviors. For example, in Portugal, you have more access to the doctor. In the US, you see a physician assistant for vitals and get five minutes with the doctor. In Portugal, you get a half hour with the doctor with no nurse or assistant.

Follow-up is not very good anywhere; nobody does much follow-up. But patient-doctor communication is higher bandwidth in Portugal because of longer appointment times. The follow-up problem exists universally but patient experience varies regionally. The US system is unique in its complexity.

I wasn't aware of that. I sprained my ankle badly in Rome and had an interesting experience. Are you primarily focusing on the US market?

Yes, the US market is where we think there's the biggest impact and opportunity to save because more dollars are spent. Healthcare is a target-rich environment for this type of work. I like to try hard things and go after them.

What is the next big thing you'd like to implement to make your product better?

We're taking the core workflow we've built and making it easier to integrate. We're building out-of-the-box connectors to get customers up and running in days or weeks instead of longer onboarding. Integration is key. Another big piece is building a compliance engine that handles all regulatory rules.

The US has a crazy hodgepodge of rules varying by state. There are standard rules about phone calls, healthcare rules varying state by state, and AI rules emerging. For example, in Utah, if you have generative AI in a scheduling conversation, you must disclose it if asked. For clinical conversations, you must get affirmative consent. Mapping this for every state and scenario is crazy, so we've built a core compliance engine around this.

How do you feel AI will impact jobs in healthcare? Will it replace anyone or just make it easier for healthcare workers to do more and provide better care?

This is a big question and one reason I like working in healthcare. There's not enough human capacity to provide care; there aren't enough doctors or nurses. I believe technology can scale care and let providers do more rather than cut the number of providers. Some places might try to cut corners, but our customers say it's about letting them scale what they can do.

I'm glad to hear that. Many people are concerned about AI replacing jobs. Some healthcare areas might be susceptible, but it's a challenge to think about. I appreciate your honesty because some people act like it won't occur. In healthcare, people are nervous about AI deciding things like claim rejections. I feel healthcare needs humans because care is human. AI can help but I get nervous about some applications allowing AI to decide access to care.

Let's talk about concerns with AI in healthcare providers and companies. Where do you think concerns lie and how AI might be used?

One challenge is the battle of AI on both sides: providers use AI to build expensive health codes, and insurance companies use AI to stop overbilling. This is a revenue battle, not about improving patients or efficiency. That side sits strangely with me. The biggest question is AI around claim processing. It's a manual process, but denials shouldn't come from a model; they need a person.

I've talked to folks in healthcare. AI is good to scale approvals but humans must be in the loop on denials or bad decisions. Responsible AI deployment means speeding things up but letting people evaluate tough cases. There are people on both sides of this, so it's important to note.

On a positive note, what is your vision for the impact your product will make in healthcare in the next 3 to 5 years?

In my perfect vision, every patient gets follow-up conversations from their providers and healthcare professionals. That doesn't happen now; most people never get followed up. Someone checks in, sees how you're doing, and helps you when you're not in a good spot.

Totally agree. I can't imagine this will be easy. What do you think is the biggest barrier to entry stopping that vision?

The biggest barrier is healthcare complexity. IT systems are fragmented; data doesn't flow from one place to another. There's no one system of record with everything. That's the technical barrier. Adoption barriers include aligning financial incentives, which exist but aren't well prescribed or easy to understand.

What is a specific thing you'd like to bring to your product that you don't have right now?

There's always a big list. For us, the biggest thing is pushing forward on real-time compliance. We have a good portion mapped out, but the next step is making it a core engine to figure out when to call people and when is the right time to contact the patient. Managing that data side is the next big piece.

Outside healthcare, what are some things you're excited about with AI in business moving forward?

I think huge changes are coming across industries. Some changes will be painful; others better. From software development, we're able to get first versions of features out much faster to validate solving the right problem. Time and resources get wasted waiting to figure out if you've solved the problem right. I'm excited about faster iteration.

Our product manager can get a prototype in front of customers in hours or days instead of waiting for development sprints, which often build something wrong the first time.

I appreciate all the insights. For a closing statement, what is your favorite AI for personal use right now?

That's tricky. I still struggle balancing tools. I've been a huge ChatGPT user from the beginning with most of my history and context in it. Google and Gemini have stepped up. I often put my biggest challenges in both and steer towards whichever gives better answers. Lately, I've gone down the Gemini path more. Gemini is incredible; it's shifted from a toy to where I end up more often.

What frustrates me is integration into my data sets. Nobody has solved how to look at my email, drive, and context to distill knowledge for decisions. I think they'll get there soon. Microsoft with Office 365 and Google are dominant spots to leverage integration, but nobody has it right yet.

I like Gemini's multimodal capabilities. I created a workflow using Gemini to QA video edits, which is crazy. It’s insane how models are improving. I had an idea to QA videos through multimodal capabilities in June but no model could do it well. Gemini's screen and video understanding benchmark is 77%, next highest 20-30%. I tweaked automation and put in new models in two hours, and my video editing team now gets automatic quality assurance based on my training.

The biggest change for me in the last months is speech integration. I used to type all my emails; now 75% of my documents are transcribed using new tools. My favorite is Monologue, which integrates well into Mac. It’s a hybrid of speech to text and app-specific context prompts using screen context for better results. It’s an amazing multimodal experience.

To check out what we're doing at Conig, visit consig.ai. Feel free to reach out to me on LinkedIn; search for RJ Burnham. Thanks for listening and watching.