Forus raises $160M at a $1B valuation to use AI to help patients access expensive medications faster

Speaker 1: Down. Completely flat. Up point 4%.

Speaker 2: Interesting. Value.

Speaker 1: Anyway, we have our next guest, Sahir from Forus. He is the cofounder the founder and CEO. Welcome to the show. How are you doing?

Speaker 6: Good, guys. How are

Speaker 1: doing? Fantastic. First time on the show. Please introduce yourself. Tell us what you're building.

Speaker 6: Yeah. Thanks for having me. So today, we are publicly announcing for the first time our company Forus, which has raised a $160,000,000 from Thrive Capital, General Catalyst, Excel.

Speaker 5: This is It's a lot out of

Speaker 2: this What a You

Speaker 1: have coming out of stealth with a $1,000,000,000 valuation. Congratulations. That's amazing.

Speaker 2: I usually give this advice to founders. Yes. Like don't just talk about what you're doing until you're a unicorn. Just focus on the basics. Don't worry about the news. Don't worry

Speaker 3: about the media.

Speaker 2: No. It's actually true. Underrated strategy if you can pull it off.

Speaker 1: Yeah. But let's talk about what you're actually building. Talk about the business, talk about the development, maybe a little bit of history that went into building the company, start deciding to start the company.

Speaker 6: Yeah. Sure. So what we do today is we help people get access to medicine faster, easier, cheaper. Most impactful are people with high cost and complex conditions. So think things like autoimmune diseases, COPD, cancers, anything where the drugs are unaffordable without insurance coverage or financial assistance Mhmm. Or have complicated supply chains. Remind to get doctor and patient weeks of phone calls and research and paperwork, get their medicine on time and affordably. Mhmm. What we're doing is using AI to take on all that complexity and extract it from them to get their medicine without any of the burden falling on their shoulders.

Speaker 1: Yep.

Speaker 6: But we do all of that entirely for free. Okay. It allows us to build a large network of doctors, patients, and the data and systems between them. Okay. And we're gonna use that network on the other side of our business to partner with life science companies. Think companies like Pfizer, Lilly, Johnson and Johnson, helping them develop and bring to market new drugs faster and more efficiently. This is really a two sided model.

Speaker 1: Okay. Two sided model. I feel like every time we just talk to health care, it's like 10 sided model with, like, insurers and health care networks and doctors, and then the the hospitals have private equity backed, and they have a different set of incentives. Like, who who are you are you cutting anyone out? Who are you not interfacing with? Are you interfacing with insurance or not? Is that a deliberate choice? Will that change over time?

Speaker 6: No. We so that's actually part of why this is such an impactful product for the core users. So we sell the doctor's offices Okay. As free product to help them automate all the complexity involved in dealing with all these other pieces. So think the insurance companies, the PBMs, the pharmacies

Speaker 1: Sure.

Speaker 6: Any other other aspects. And, really, the goal is to automate every process that involves them working with those folks externally so they can focus on treating their patients and not have to worry about all that complexity underneath the hood. We interact with all those folks to kinda push things to the system, but it puts us in a position where we are kind of at the center of each individual transaction and have kind of a unique view into what's going on. And that's what sets us up then partner with these life science and biopharma companies Mhmm. To help them in a unique way on bringing to market new therapies.

Speaker 1: Have you developed fax machine super intelligence yet?

Speaker 6: I would bet that we have some of the best fax AI

Speaker 1: No way. And that's both on is that both on sending and then receiving and transcribing, OCR ing, and understanding and then putting in some sort of database? So, like, do are you bidirectional in your faxing?

Speaker 6: We are bidirectional. There's, like, a pretty decent chunk in transactions that can only go through fax.

Speaker 1: Yeah.

Speaker 6: In reality, fax is actually a, considered HIPAA compliant.

Speaker 7: Yeah.

Speaker 6: So you have a lot of leeway there as opposed to in other transactions, but it's also very reliable. If you can't get someone on the phone, you don't have their contact information, you can almost always find their fax in publicly. So despite the fact that it seems insane, it's a shockingly reliable form of communication in the system.

Speaker 1: Yeah. And so what else is is what else is enabled or accelerated with AI? Because I can imagine you can use AI to build, you know, business logic and software. Like, you can build SaaS faster faster for deterministic workflows. But then you can also do sort of agentic things. When I hear about like finding someone's fax number online, that sounds like, yeah, you could Google it, but that's usually a person. There's not really an API, but with an agent, there sort of becomes an API. And so where where are you doing agentic tasks on an ongoing basis versus using AI to develop sort of SaaS?

Speaker 6: Yeah. So there's really two layers we think about Yeah. And what makes the problems hard. The first is kinda what you're describing, a genuinely hard set of agent problems. You're navigating this multistep path dependent process where there's almost never an obvious correct answer. Mhmm. And so it's usually not even cataloged anywhere what you're supposed to do or what the right answer is. You're working with imperfect information every term. What insurance does someone have? What does it cover? What pharmacies work for this specific drug? What approval criteria exist? What medications have this person tried before? And you're interacting with all of that, like you mentioned, through faxes, phone calls, poorly designed websites, and scattered information.

Speaker 4: Mhmm.

Speaker 6: And so that's kind of one huge piece of it. The second is a set of really complex ML problems where you're using dozens, sometimes hundreds of pages of medical records per person to answer very precise questions about someone's medical history across thousands of medications and diagnoses, dozens of subspecialties, and tens of thousands of doctors who all document and write notes about their patients and diagnose them and treat them differently. And so that same capability is what powers everything from navigating insurance authorizations to identifying whether a patient might be eligible for clinical research. That system kinda never stops evolving. Right? New drugs, new guidelines, new insurance policies. And so as we're building and scaling, the complexity keeps increasing, and you're kinda constantly regionalizing everything you've built in your platform.

Speaker 1: Mhmm. Last question for me. Talk about the data side of the business. I imagine that there's you want every doctor's fax machine number. That's a data problem. You might be able to scrape it. You might be able to buy it. There's also the data that goes into maybe fine tuning models, maybe just setting up models for success, giving them the correct context. Like, what is the shape of the data problem and how is it evolving?

Speaker 6: Yeah. I mean, it's kind of like what you described. It's through every single step in the process. Mhmm. You can imagine where we sit, we see this from end to end. Right? So we see everything from the clinical information, how the doctor made their diagnostic decision, all the way through to logistically what got stuck in the system and prevented someone from getting their medication or results from it being late. All of that's really relevant in understanding not just a, how do we translate that into helping us automate more effectively and have more predictability in every step of the process, but also in then informing decisions that our partners are making on how to kinda move therapies forward in the research and development process, how to think about launching them, where to allocate their resources so those medications can reach as many people as possible.

Speaker 1: Got it.

Speaker 2: How big can the business be?

Speaker 6: So and there's two parts. Right? We think this can be the most important company in life sciences. Our goal number one is ubiquity. Right? Every doctor's office and patient in the country should be using our product. That position doesn't obviously just let us eliminate friction for more people, but it puts us in a position to accelerate every step of bringing new therapies to market on the other side. Right? When a new molecule is ready

Speaker 2: to test Do you become basically just like a is the legacy, like, legacy version of this and maybe in another industry like a distributor? Is that like the less

Speaker 6: Distributors are like a partial analogy, but, like, a lot of what we're gonna do is not the actual physical supply chain.

Speaker 5: We're

Speaker 6: really helping on all kind of decision and actions around that. So think about when you have a new drug. Right? There's a ton of stuff happening right now in ad drug discovery. All of that is translating into more and more hypothetical molecules. But how do we translate those molecules into production, real life mass market medicines? Mhmm. From the moment a new molecule is ready to test in people all the way through all those steps, you have to figure out, you know, how do you identify the right clinical trial sites? Which patients do you recruit are not not contaminated by other medications? When a drug launches, which doctors are the most patients that look like the patients that did best in the phase three trials and therefore the right, you know, early adopters and can benefit the most? As it scales, how do you kind of use real world usage to identify where there are new conditions, the same drug can be effective, or there are gaps in the market? That full loop from clinical development through mass market access hasn't really been possible before, and we think we can kinda be the main chassis through which that happens. Today, we are, as you can imagine, kind of growing in a ton of different specialties. And the kind of first specialty we launched, we're already kind of approaching a third of the country. So you can imagine, we really can get really far along and become the best partner to facilitate all these decisions.

Speaker 1: Well, congratulations on the progress, and congratulations on the round. Thank you so much for coming on the shop.

Speaker 2: Great Thanks for having me. Day of media. Yeah. What were you doing before this?

Speaker 6: So I used to work at this company called Oscar, you're familiar with it.

Speaker 5: Was that

Speaker 1: That makes so much sense. I thinking this whole time.

Speaker 2: Yeah. I was thinking this whole time. Yeah. Must have been an Oscar. Thrive universe. Yep. Thrive universe. Thrive universe. Cinematic universe. Great wine. Great wine.

Speaker 1: You and the SF Giants, similar portfolios. Fantastic. We'll talk to you soon. Goodbye.

Speaker 2: Cheers. Thanks, guys.

Speaker 1: Take care. Up next.