Kyle Kiser is Using AI to Make Your Patient Experience Better (EP 140)

The Harry Glorikian Show 

We’ve had it all happen. You leave a doctor’s office with a prescription, and you get to the pharmacy, and then you hit a wall. The medication isn’t covered, or it requires pre-authorization, or the out-of-pocket cost is just shocking. It’s a breakdown that happens at the intersection of healthcare, technology, and policy. And it’s a breakdown that costs patients dearly, sometimes literally. Today’s guest is someone who’s been working to fix that. Kyle Kaiser is the CEO of Arrive Health, formerly known as RXRevu, and his company is on a mission to close the information gap between providers, payers, and patients. Under his leadership, Arrive Health has become a major force in embedding real-time prescription pricing and coverage data directly into the clinical workflow, so doctors can make informed prescribing decisions at the moment it matters most.

Here’s our conversation:

Harry Glorikian: Kyle, welcome back to the show.

Kyle Kiser: Glad to be here.

Harry Glorikian: So Kyle, I was thinking about this, and I was like, well, we did the recording last time, and I thought, okay, there might be some people who didn’t hear the first version. So I want to rewind a little bit. Arrive Health now, which was originally RXRevu, grew out of a pretty personal story, if I remember correctly. Your co-founder, Dr. Kevin O’Brien, was trying to help his mom afford her meds. Can you take us back to that moment and share how it evolved into, “let’s start a company and solve this problem”? That’s a bit of a jump, but help me get there and help our audience get there.

Kyle Kiser: Yeah, happy to. That is one of my favorite stories. This all started out of that outreach. Kevin, who’s still a doctor in Denver and still involved with the business, had his mom come to him with a high out-of-pocket cost. And like any good son with the skills to help, he helped. He looked at her medications and found there were opportunities to save. These were low-hanging fruit type things. For example, she was taking a branded medication that was essentially made up of two generic drugs, so he advised her to take the generics instead. Or he found therapeutic equivalents that were better aligned with her insurance. He ended up cutting her costs in half.

What that experience showed him was that his patients were dealing with the same problem inside his clinic. So Kevin started accumulating all the ways he was identifying to save money on medications. He built a massive, unwieldy spreadsheet based on payer data. Originally, he thought he’d self-publish it as a book on Amazon, just to get it out in the world. But the problem was, that book would have been obsolete the next day. Drug prices change so frequently, and plan designs are so dynamic and complicated.

Around that time, he met some of the other co-founders and convinced them this wasn’t a book — it was an API business. And that’s how the whole idea came together. That was the launch of what is now Arrive Health. And it still lives and breathes today. We have a company mantra called “Lucy Up.” It’s our way of saying to each other that this is about something bigger than our own success and ambition. Kevin’s mom is named Lucy, and this mantra is a daily reminder of the human reason we do this work.

Harry Glorikian: That’s powerful. And what it taught you early on is that the provider is the right steward of that information, right? The clinical decision happens between the patient and provider.

Kyle Kiser: Exactly. Years ago, and not even that many years ago, when deductibles were in the hundreds of dollars and copays were in the tens, cost didn’t have to be part of the clinical visit. But now, more than half of America has multi-thousand-dollar deductibles. Out-of-pocket costs are often in the hundreds or thousands of dollars. So knowing whether a patient can afford their medication, and whether it aligns with their insurance, is now a core part of delivering great care.

From that origin, we’ve always pursued providers as strategic partners. Funny enough, that was actually the Chief Medical Officer of UCHealth calling me in the middle of a speech. UCHealth is where we were incubated. It was crucial for us to work directly with providers, because it taught us how to introduce this information in a way that’s easier, not harder — fewer clicks, not more — and trustworthy, despite the complexity.

That’s the only way providers will engage. None of this would have happened if we hadn’t come at the problem from the provider’s point of view. The reason this problem hadn’t been solved before is that it was always outside-in. Health plans and PBMs were saying, “Here’s how we think this should work.” Not physicians saying, “Here’s how we’ll actually engage with it.” That difference in approach is what allowed us to rebuild the market-leading network we have today.

Harry Glorikian: I mean, when you think about it, it’s fascinating. One story captures a universal frustration. And you guys seem to have found a wedge, a way in to change a system that, if you talk to the average person, seems immovable. If you talk to my wife, she’s like, “Okay, it’s my day to spend two or three hours going through insurance bills, making phone calls, just to make sure everything went through.” It’s a universal frustration, and pharmaceutical expenses are a huge problem for most people. If your system can help alleviate that, I think that’s a huge win.

So last year, if I remember correctly, you guys rebranded into Arrive Health. And that wasn’t just a name change, right? I think it signaled something deeper. What were you trying to reflect with that shift, and how has the mission evolved?

Kyle Kiser: Yeah, I think our goal and our vision for the business for a long time has been improving the value of healthcare through better decision-making at the point of care. And that future inevitably evolves beyond pharmacy. Pharmacy is still our primary focus — we’re doing a few hundred million transactions a year now — but there are a billion or more new prescription decisions happening every year. There’s still plenty of opportunity to make that better.

Not to mention things like prior authorization and affordability more broadly. But the world is evolving. The processes that people face around affordability and access are going to become universal, and they’ll increasingly be brought into the provider workflow. Decision-making around clinical best practices — and how that aligns or doesn’t align with payment — is becoming part of the clinical visit. Through advancements in technology, like generative AI and other tools, prior authorization will become available at the point of care and automatable from there. That’s the future we’re preparing for.

So the name Arrive Health positions us really well for that. It’s definitely something we’re thinking about every day.

Harry Glorikian: Yeah, I was thinking about the word “arrive” as a verb — like arriving in a timely and affordable way at the moment care is needed. But I want to get a bit more technical here. One of your signature offerings is the ability to show real-time, patient-specific drug costs right inside the provider’s electronic health record, the EHR. For someone listening, that might sound simple, like, “Why isn’t that already happening?” But I imagine the execution is quite complex. Can you walk us through what happens under the hood when a doctor prescribes a drug using your system?

Kyle Kiser: Sure. Let me start with what happens more often — without us. We’re probably approaching the point where we can impact about half of all prescription decisions in one way or another, but there are still plenty we don’t touch. Before we existed, here’s how it typically worked: a doctor makes a prescribing decision, routes the script, and the first experience the patient has at the pharmacy is often a “no.”

They show up, and it’s, “You need to complete a prior authorization,” or, “This will be $2,000,” and they can’t afford it. Or, “You can’t fill it here, you have to go to a different pharmacy.” The whole thing is full of redirections and rejections.

So, put simply, what we do is take that experience away. We give providers what’s essentially the teacher’s guide to the test. We say, “You were going to prescribe Drug A, but Drug B is the insurance-preferred alternative. Is that okay with you?” Or, “You were going to choose Pharmacy A, but Pharmacy B is in-network. Is that okay?” We surface that information at the moment of prescribing — inside the EHR — so the provider doesn’t have to use a separate system or swivel to another screen. It’s part of their normal workflow.

Harry Glorikian: That’s awesome. And that cost information — it’s not just “how much is the drug at Walgreens,” right? It includes nuances like the patient’s insurance plan, deductibles, prior auth requirements, quantity limits. If I understood it correctly, it’s quite a bit more comprehensive.

Kyle Kiser: That’s right. And if you really want to get into the weeds, what’s required to make that work — especially in the more expensive or complex cases — is that you have to bring the intelligence of what the pharmacist was doing at the point of sale into a sub-second transaction at the point of care.

One of the most valuable and underappreciated members of the care team is the pharmacist. They’re like the left tackle — super important, but you rarely hear about them. At the point of sale, a pharmacist will see, “This NDC doesn’t work, I’ll switch it to this one, and now it’s covered.” All that intelligence, that expertise, we’ve built into our tool as a learning model.

On top of that, every EMR — every electronic health record system — structures its drug data a little differently. We’ve had to build technology that can translate those variations into the correct type of request to the health plan, and then translate the response back into something the EMR can understand and display.

The result? We get a really high success rate. We’re able to provide a price more often, providers trust the information more, and they engage more. And when they engage more, they’re more likely to switch to those lower-cost options. But it all starts with getting that first step right — delivering accurate, timely, actionable data.

Harry Glorikian: Yeah, I mean, that’s one of the things I was thinking about when I was prepping for this — you’re not just showing a cost. You’re showing options: maybe a cheaper pharmacy, a lower-cost alternative, or even a different dosage that’s more affordable. Have you seen that lead to prescribing changes?

Kyle Kiser: We have. We actually published a white paper on this, maybe more than a year ago now. We found that almost 50% of the time — I think the number was 48%, though I’m not sure I can quote it exactly — a provider prescribed something that cost more than $50 when a $0 option was available. Which is wild.

It’s not because the providers didn’t care. They weren’t making bad decisions on purpose. It was a visibility problem. Inside their workflow, they might be used to prescribing a particular thing. They’ve got order sets and muscle memory. So our job is to bring that cost and coverage information into the workflow, at the right moment, and give them the opportunity to make a better-informed decision.

Harry Glorikian: Yeah. If I asked my doctor — who’s also my friend — about this, he’d say, “The fact that I even have to think about this is insane.” But what you’re doing is giving them a real-time decision engine. It’s part clinician, part case manager, part shopper. It’s a shift in mindset because of how the system is set up.

Kyle Kiser: Exactly right. Think about a typical clinical day. A doctor sees 40 patients in 15-minute increments. Those 40 patients may represent 12 or 15 different insurance configurations. Even patients with the same insurer — say Blue Cross Blue Shield — can have completely different formularies, pharmacy networks, and cost structures depending on whether it’s Medicare, Medicaid, or a self-funded employer plan.

It’s unreasonable to expect any human to keep all of that in their head. That’s a machine problem. A machine can solve it. So we’ve approached it by saying: only tell the provider exactly what they need, when they need it, and then get out of the way.

When we rebranded to Arrive Health, the tagline we chose was “clearing the way for better health.” And that’s what we think technology can do — get out of the way of the doctor-patient relationship. That’s the most trusted relationship in healthcare. People call their doctor because they want an answer. They tolerate calls from their health plan. So we empower doctors with the right information so they can give the right answers.

Harry Glorikian: Well, and it alleviates a lot of frustration too. I mean, I can just imagine a patient saying, “Why did you prescribe me a $2,000 drug?” And the provider replying, “I didn’t know it was $2,000.” But your platform gives them that information up front. Let’s talk about one of the big elephants in the room — prior authorization. Every provider hates it. Every patient who’s familiar with it dreads it. But it’s deeply entrenched in the system. How are you tackling that piece?

Kyle Kiser: In a couple of ways. The industry has a bit of a Sophie’s Choice when it comes to prior auth. You can’t eliminate it entirely because it plays an important role in managing cost. It’s a check and balance between clinical decision-making and financial stewardship. Maybe one day we’ll have a better tool for that purpose, but for now, prior auth is what we’ve got.

So the question becomes, how do we make it more efficient? First, our real-time benefit technology determines whether prior auth is required. That’s a personalized insight — specific to the patient — whereas before, it was a general rule at the group or plan level. So historically, the EHR wouldn’t know if prior auth had already been satisfied. Now, we can say in real time, “That’s already been done, you’re good to go.”

Second, how do we initiate the process earlier? Even today, most electronic prior auth happens after a patient experiences a denial. That means we’re still starting from “no.” We’re working to change that — we want to initiate prior auth at the point of prescription, with the goal of resolving it before the patient ever gets to the pharmacy counter. So by the time they arrive, all the hurdles have been cleared.

That’s the same philosophical approach we’ve taken with cost: remove the friction before it hits the patient. Prior auth requires more complex data exchange between the EMR and the payer, but it’s the same idea.

Harry Glorikian: You mentioned earlier someone from UCHealth calling you. If I read correctly, you’ve actually streamlined their prior auth process using this real-time intelligence. So what you just described — that’s already in place with them?

Kyle Kiser: It is. We’ve implemented this both at UCHealth and at UPMC in Pittsburgh — two really innovative systems. One of the most recent things we did at UPMC was take all their prior authorization policies and train a model on them. Then we turned that into a usable summary — kind of like the teacher’s guide again — and surfaced it at the beginning of the prescribing process.

So, say you want to prescribe a GLP-1. We’ll say, “Here’s what you need — a diagnosis of diabetes, a BMI over 28, an A1C above 7.5, and that A1C test must be from the last six months. Do you have all that?” What we’ve seen is pretty astonishing. We saw a 50% reduction in denials — not necessarily more approvals, but fewer denials.

That’s the Sophie’s Choice again — we’re getting the right people on the medication, and we’re still using prior auth where it’s helpful, like preventing inappropriate or cost-ineffective use. So to me, that’s where the opportunity lies. How do you get that information into the hands of the decision-maker at the start of the process?

Then, if the decision requires a prior auth, we create automation around it. We do the data exchange and solve the problem before the patient hears “no.” That’s scalable today. We can take policies from any health plan and do what we did with UPMC — embed them in workflow without needing to change legacy systems.

That’s the big limitation in healthcare — the old, complex infrastructure. But in this case, we can work within existing systems and still improve the process. And I think that’s one of our best opportunities.

Harry Glorikian: Yeah, I was going to ask, how are the new generative AI approaches changing what you’re able to do? Two or three years ago, what you just described wasn’t possible. But now, there are so many new capabilities. How is that impacting the business?

Kyle Kiser: I don’t think it’s just us — the whole industry is being transformed. A big limitation in cutting administrative waste has been that much of the necessary data is unstructured. The very first product we built almost ten years ago combined clinical best practices and cost information into a one-click ordering experience. The idea was, if we surface the right recommendation at the point of care, will the provider choose it?

And they did — two-thirds of the time, which is an incredible rate for point-of-care technology. But the challenge back then was twofold. First, how do you get clinical best practices into a structured form? These are 60-page documents — probably the only person who’s read them in full is the one who wrote them. Second, how do you extract the right information from the EHR when it often lives in a physician’s note instead of a structured field?

Now, both ends of that problem have been largely solved. That opens up huge potential — and it’s exactly what we’re pursuing. These capabilities will show up first in things like revenue cycle and prior authorization, where the clinical risk is low but the administrative burden is high. We’re not trying to replace clinical judgment, we’re just clearing out the noise.

Harry Glorikian: Zooming out a little, I want to get a sense of scale. How many providers use the tool? What outcomes are you seeing?

Kyle Kiser: We focus a lot on behavior change — how often do we motivate a switch to a lower-cost option? When we launched our benefit check product in earnest in 2019, we were doing a few thousand transactions a month. This month, we’ll do a little over 25 million. That’s a good growth curve, and it’s still accelerating.

One thing that’s always stuck with me — I read years ago in Health Affairs that one of the most requested data points by ambulatory physicians was the cost of medications. They want this information. Now, we’re finally able to deliver it.

In terms of behavior change, generic dispense rates are already high because formularies have trained physicians to pick generics over brands. But even within generics, there’s still a lot of variability. So picking the right generic matters. When we present alternatives, we see a pretty healthy switch rate — probably in the 20% range. That’s huge. Most provider behavior change tools see single-digit rates. And based on the data we’ve seen, our tool is trusted about twice as often as any of the others.

Harry Glorikian: It’s funny — if you’ve got an older doctor, change might be harder. But if someone has been using your system as their career develops, I imagine they’re more likely to trust it.

Kyle Kiser: Yep, that’s right.

Harry Glorikian: What about payers and PBMs? Are they being collaborative? Because part of me wants to say they’re gatekeepers more than collaborators.

Kyle Kiser: No, they’re absolutely shoulder to shoulder with us. They’re our data sources. We rely on their systems to deliver real-time cost and coverage information. Often, they’re the ones providing the preferred alternatives. So we have to directly integrate with their systems to do what we do.

There’s been a lot of progress. A couple years ago, CMS passed a rule requiring at least Part D plans to offer this kind of capability. That brought more people to the table. And payers want to solve this too. They experience massive costs from administering prior auth. They’re stuck between a rock and a hard place — trying to reduce prior auth burdens without letting costs spiral out of control.

Since we started working on this, drug costs have grown from the low teens to the low 20s as a percentage of total healthcare spending. So this is a critical problem for them as well, and our tool is an important part of the solution.

Harry Glorikian: If you had generative AI ten years ago, you might’ve moved faster. I think I read that you partnered with AWS on an AI-powered solution — is that different from what you’ve already been describing?

Kyle Kiser: That’s our PA project. We took those massive prior authorization policy documents and turned them into decision support tools powered by generative AI, surfaced right at the point of prescribing. AWS collaborated with us and UPMC to launch that. And to your point — it moved fast. From a cold start to being live across the UPMC network in less than nine months.

Harry Glorikian: That’s incredible. Usually, getting something like that deployed takes years. These new tools really do compress the development cycle.

Kyle Kiser: That’s been our experience too. We’re seeing things move much faster.

Harry Glorikian: And providers are accepting it, which is the other part.

Kyle Kiser: Yes. At least in the administrative areas of healthcare — which is where we focus — there’s high demand for these solutions. The current processes are still based on phone calls, faxes, and manual effort. That’s frustrating for patients and costly for the system. These are solvable problems, and they’re a great use case for this kind of technology.

Harry Glorikian: And I imagine this reduces stress for physicians too.

Kyle Kiser: Absolutely. One stat I keep in my head — 65 million calls go back to physicians’ offices every year just about medication issues. Most of those are simple things: wrong med, wrong quantity, wrong pharmacy. We can make those disappear.

Harry Glorikian: It’s interesting — usually on this show I’m talking about AI and drug discovery. But now we’re talking about drug delivery — getting the right patient the right medication faster. A quick side question: your business model is interesting. Risk-bearing entities like payers and PBMs foot the bill. Why is that alignment important for sustainability?

Kyle Kiser: Ultimately, that’s where the benefit accrues. Whoever pays for the medication should benefit from the better decision. So we’ve aligned our business model with that.

We typically work with PBMs because they’re the most aggregated version of the risk-bearing entity. But it flows all the way through — the PBM works for the health plan, the health plan works for the plan sponsor, and so on.

It also makes sense from a provider adoption perspective. You don’t want to create cost barriers for provider organizations to adopt the tool. We’re already asking them to implement it, which takes effort. So we structured this like a typical multi-sided network business. Think of digital ad exchanges — there’s supply (pricing data) and demand (provider usage), and we’re the network that gets the right information to the right user.

Harry Glorikian: But it’s complicated, right? Each stakeholder has different incentives. How do you make it all work?

Kyle Kiser: That’s true. I’ve spent most of my career working in health tech focused on payers, and usually there are major misalignments between stakeholders. But I think this problem — medication access — is one of the rare areas where incentives actually align.

Risk-bearing entities want to manage costs, providers want to help patients, and patients want to get on their meds without friction. If a provider is in a value-based arrangement, they’re incentivized to manage chronic disease — and that means adherence. Even pharma wins when the right patients get on the right drugs faster. So this is a high-alignment use case, which makes it special.

Harry Glorikian: So that brings us to your expansion beyond meds — into lab and radiology pricing, maybe even procedure-level transparency. What does that expansion look like?

Kyle Kiser: That’s not something we’re immediately focused on. We see that future, and we’re prepared to move toward it, but for now, our focus is still the pharma value chain. The drug side is better structured from a data perspective, so the problems are more solvable. The medical benefits side is still much more unstructured — that’ll change over the next three to five years. But in the meantime, there’s still a ton of opportunity on the pharmacy side.

Harry Glorikian: I imagine that recent policy — like the No Surprises Act and the CMS transparency rules — has been helpful in moving things forward.

Kyle Kiser: Definitely. Those policies work together — payer transparency, provider transparency, No Surprises — they’re like a triple force pushing the system toward greater visibility. That will absolutely improve the patient experience.

Harry Glorikian: One thing we talked about last time was how this isn’t just about price transparency. It’s really about access to care. That feels like a subtle but powerful reframe. How do you think about that internally?

Kyle Kiser: One of the best ways to frame it is “time to therapy” — how long it takes from prescription to the medication being in a patient’s hands. In some best-case scenarios, we’ve taken that time from days to zero. In more complex prior auth cases, we’ve gone from three or four days to three or four hours.

That’s what access really means. Once it’s the right med, how quickly can we remove the friction? If the health plan has the info it needs, the provider doesn’t have to do extra work, the patient gets the drug, and everyone wins. That’s our North Star.

Harry Glorikian: I don’t think my health system uses your platform — I haven’t had that experience. But looking ahead, what are you most excited about over the next 12 to 24 months?

Kyle Kiser: We’re excited to advance the prior auth work and get more health plans on board with the new generative AI model. We’re also expanding how we support affordability more broadly — including cash pay options, discount cards, and other payment methods — and making those available at the point of care.

We’re also building capabilities to extend this experience directly to patients. We’re working with a few EMRs to create a shoppable experience inside patient portals. That way, once a medication is prescribed, it shows up in the app and the patient can make a more informed choice. Those tools are already well-adopted, so this has huge potential.

Harry Glorikian: That would be awesome. I’ve seen situations where the cash pay option is cheaper than the insurance one. So having that option would be huge.

Kyle Kiser: Yeah, and I hope through that process we bring pharmacies to the table. The payment ecosystem has made it really hard for them. Sometimes they’re locked into rates that don’t work for them financially, or they’re squeezed by thin or even negative margins.

But pharmacists are essential members of the care team. My mom literally brings Christmas gifts to the Walgreens staff — she sees them more than almost anyone. That relationship matters. We want pharmacies to come along on this journey in a sustainable way.

Harry Glorikian: You’ve been doing this for a long time. If you were advising other founders or innovators in healthcare — which, let’s face it, is a very messy space — what’s one piece of hard-earned wisdom you’d pass along?

Kyle Kiser: That’s a great question. For me, it comes back to the “Lucy Up” story. It’s always going to take longer than you think. It’s always going to be harder than you expect. And even once it starts to feel easier, it won’t last — because as you get better, the problems get harder.

The only thing that builds the resilience to keep going is an intense focus on why the work matters. That’s what I’m most proud of in our business and our team. We never lose sight of that.

It’s not leadership telling the team to care — it’s the team holding each other accountable. When we go to build new things, someone will ask, “Does this prioritize the patient?” That’s one of our core values.

If you stay organized around purpose — and genuinely care — you’ll make better decisions, build a stronger culture, and move faster. And ultimately, you’ll build something more durable. People wake up and think, “It shouldn’t be this way. Patients shouldn’t experience a ‘no’ first.” And that clarity helps us keep chipping away at those barriers.

Harry Glorikian: I tell everyone who comes on the show — move faster, because I’m getting older and I want to see these problems solved!

Kyle Kiser: Plenty of wood to chop.

Harry Glorikian: Great having you on the show again, Kyle. I’m really excited about the progress Arrive Health has made, and I wish you continued success.

Kyle Kiser: Thanks, Harry. It’s great to be back.