đ§đ Healthcare Skipped Three Platform Shifts and Went Straight to AI | Alamin Uddin, NexHealth
Why 75% of dentists still have a server in the closet, and how NexHealth went from $4k in the bank and a maxed-out Amex to powering 81% of new healthcare startups
Today, NexHealthâs healthcare infrastructure serves 89 million patients. But just a few years after starting the company, they had $4,000 in the bank, a maxed-out Amex card, and were one month from running out of money.
Fast forward to 2026, I talked to NexHealth co-founder and CEO Alamin Uddin about why healthcare skipped the internet, cloud, and mobile and went straight to AI, the reason healthcare has no universal API, funding Nexhealth with side hustles, the $36k customer pre-pay that saved the business, how 81% of new AI healthcare startups are built on NexHealth today, where he thinks AI value actually accrues in healthcare over the long-term, and building a company to outlast OpenAI.
Support this Episodeâs Sponsors
Numeral: The end-to-end platform for sales tax and compliance.
Flex: Sign-up for Flex Elite with code TURNER, get $1,000 here.
Amplitude: AI analytics, all you have to do is ask.
Merge: Every model. One API. Total control. Check out Mergeâs Agent Handler.
Monaco: The revenue engine for startups.
Inquire about sponsoring future episodes here.
đ Stream on Spotify and Apple
Timestamps to jump in:
0:00 Healthcare skipped 3 platform shifts and went straight to AI
4:10 75% of dentists still have on-prem servers
10:05 How data interoperability holds back healthcare innovation
16:02 Why everyone blames Epic
20:15 Building the developer platform for healthcare
24:36 Why everyone fails to fix the problem
29:11 Fragmented markets enabled developer platforms
32:32 Working as a receptionist at a doctorâs office
36:13 Building a prototype on Twilio
39:15 How incumbents went from blocking to partnering
46:05 Canvassing Soho dentists door-to-door
53:47 Reverse-engineering 40-year old databases
56:21 Funding NexHealth with side hustles for two years
57:30 The scheduling wedge no one could match
1:02:20 Raising $391k from professors and customers
1:03:46 Running out of cash, why customers kept churning
1:07:45 $4,000 in the bank and a maxed-out Amex
1:11:12 The $36k pre-pay that saved the company
1:13:24 NexHealthâs three businesses today
1:20:18 Payments and the âadmin-dayâ problem
1:26:15 72% sales win rate
1:28:27 The term sheet signed the week before COVID
1:30:26 Spending half the Series A on an acquisition
1:33:47 Raising $176M they didn't need
1:37:28 Why starting before 2022 is an advantage
1:42:22 Where AI value accrues: chips, models, the action layer
1:44:55 81% of AI products are built on NexHealth
1:48:24 Staying patient for three years after ChatGPT
1:51:25 Competitors building on their API
1:54:02 âWeâre a tech company, not healthcare companyâ
1:56:09 Hiring from outside healthcare
1:58:34 Shoes, email over Slack
2:01:21 What AI changed inside the company
2:04:16 Inspiration from Microsoft in 1977â1990
Referenced:
Find Al on X / Twitter and LinkedIn
đ Stream on YouTube, Spotify, and Apple
Transcript
Find transcripts of all prior episodes here.
Turner Novak:
Al, welcome to the show.
Alamin Uddin:
Yeah, thank you for having me. Excited.
Turner Novak:
You were just telling me about this before we started. How do you think COVID changed the healthcare industry?
Alamin Uddin:
One of the more fascinating things about COVID is this. We started the company before COVID, in 2017, and around that time we had maybe 10 or so customers.
At that time, we would get customer calls like, âHey, the guy thatâs booking my appointments, could you please tell him to change the color to yellow or red or whatever?â As in, they thought it was an actual human booking appointments in their system, not automated software.
Turner Novak:
They didnât even realize you had made software to automate it.
Alamin Uddin:
Exactly. The end users were that far from being tech-savvy.
Turner Novak:
Wow.
Alamin Uddin:
Then COVID hits, and all of a sudden they not only become tech-savvy, they actually care about amazing UX, and things being on the cloud and mobile and easy to use.
And then right after, you get AI. With AI we have this very same ICP, the one thatâs not very technical, but their tech literacy has changed so much that they can take our API and build their own solutions.
Turner Novak:
Really?
Alamin Uddin:
Yeah. From COVID to today, what is that, roughly five, six years? Depending on how long you count COVID as lasting.
Turner Novak:
Sometimes I feel like COVID started yesterday, honestly. Itâs one of those things where you go, âHoly shit, itâs almost 2027 now.â
Alamin Uddin:
Obviously a lot of bad things happened to the world, especially in healthcare.
Turner Novak:
Oh, yeah.
Alamin Uddin:
But at the same time, some of the positives: the industry probably went through 30 years of transformation. In just five years or so.
Turner Novak:
When you say the industry, you work with smaller, independent-ish local doctors. When you think of a local business like a dentist, and I think you started with dentists, who is your customer? What do they look like?
Alamin Uddin:
We have three types of customer profiles. The first and primary one, our initial wedge, is your local dental office. As of today itâs your local dental office, your local dermatology office, and so on.
Along with that, there are the private equity roll-ups. You can think of KKR coming in and rolling up the largest dental roll-up, which is owned by them, 2,000 offices and so on. Thatâs the second type of customer we have.
The third and final one is developers and innovators in the space, who use our API to build solutions that can actually action data in the real world. Meaning, if youâre a YC startup building an AI phone agent and selling it to practices, then whenever you as a consumer call that practice and the AI agent picks up, that agent needs to actually know information about you as a consumer.
Turner Novak:
Yep.
Alamin Uddin:
And NexHealth is the connector, very similar to Plaid, where these developers are building on top. Our own SaaS product is built on top of that same API. So generally we have these three types of customers: private equity roll-ups, independent offices, and developers that build for them.
Turner Novak:
You were telling me thereâs something like 700,000 independent practices across the country, and the numberâs 80% of them still have, or had until recently, literal servers in the closet to power the business. Why was that the case?
Alamin Uddin:
Today, if you look at the dental industry, 75% is still on-prem.
Turner Novak:
Really?
Alamin Uddin:
If you look at the medical industry, Iâd say 50% is on-prem. Which is dramatically different than it used to be. Five years ago, the majority, 80 to 90%, was on-prem.
Turner Novak:
Okay.
Alamin Uddin:
You go to your local dental office and thereâs a high chance they actually have a server somewhere in the office. The desktop computers theyâre using are a server-client setup, with IT managing it to this day.
The why behind it is twofold. One is legacy: a lot of these offices digitized in the early 2000s or the â90s, and at that time those were the options available to them. Modernizing that same software doesnât generally happen, for two reasons. One is change management. Migrating your system of record is really high risk for a lot of these offices.
Turner Novak:
Why is it so hard? These are basically like a doctor who got a masterâs degree, studied medicine for 10 years, and now has to upgrade to the cloud. Is it just not really an option?
Alamin Uddin:
Itâs twofold. One is that most of these incumbents actually make it hard for you to migrate in the first place.
Turner Novak:
Really?
Alamin Uddin:
Yeah. If youâre migrating, you have this SQL database in the office that you own and manage. And youâre not a tech company. Why do you even have your own server in the first place?
Turner Novak:
But thatâs a doctor. Heâs seeing patients, and heâs also managing a SQL server.
Alamin Uddin:
Exactly. And then itâs managed by a third party. You have these records stored in a certain format. So letâs say youâre going from software A to software B. The migration and the format of that data matters, and you canât lose that data either.
Itâs patient data. You have your billing data, a lot of compliance needs behind it, and of course youâre serving patients and their records. So any loss of that data, or it not being compatible with the next system youâre going with, is incredibly risky to them.
Turner Novak:
Really?
Alamin Uddin:
Most offices would rather not even take that risk.
Turner Novak:
Canât you just back it up or something? Is it all that hard?
Alamin Uddin:
Thatâs possible, but hereâs the thing. A lot of things we expect in tech, take Salesforce as our system of record, we expect Salesforce to provide us logging data or SSO and so on. In our space, at least historically, the vendors just donât provide those tools, either because of self-interest or just a bad product.
As a result, over time the providers themselves are stuck on older systems and unable to upgrade. And even when they try, they have to go through a ton of friction to do it. Mostly because the incumbents are very incentivized to keep you within their existing ecosystem, unless theyâre now selling you a cloud version themselves.
Turner Novak:
Yeah.
Alamin Uddin:
We can get into why that is. But generally, most electronic health record systems, their business model, except for a few like Athenahealth which would be an exception, is really oriented around making sure that migration of data, change management, or access to the officeâs data needs permission from the vendor themselves. For any office or any third party to be able to innovate or adopt any tools.
Turner Novak:
So how does that impact the rest of the industry?
Alamin Uddin:
Healthcare data interoperability is one of the hot-button topics of healthcare, and it has been since the â90s. The fascinating thing about this problem is that every decade or so, the federal government tries to pass some new regulation. It was actually President Clinton first that passed regulation for open healthcare data. Then Obama tried, and then Trump passed two different regulations around it, in his first term and his second.
Turner Novak:
Now he did it again.
Alamin Uddin:
Exactly. So itâs a big problem within the space. What it basically means is this. You and I go to doctorâs offices. The majority of us were born in a hospital, and throughout our lives weâve gone to many different doctors, and as a consumer youâve generated all this health data.
If I were to ask you today, âTurner, whereâs XYZ prescription, whereâs your health data?â, the more general answer youâd give is, âWell, I donât really know,â or, âYou know what, this is the doctor I go to, maybe they have it.â And if I asked, âHow many healthcare apps do you have on your phone?â, most consumers would probably just have Apple Health. Versus the fintech space: if I went through your phone, you probably have four to five fintech apps.
Turner Novak:
Oh, I probably have more than that.
Alamin Uddin:
Right, obviously. So thereâs a ton of innovation. Where the healthcare space is today is probably where fintech was 10, 15 years ago. Thereâs a ton of innovation happening for consumers and businesses on the fintech side, but not the healthcare side.
The reason is the data interoperability problem, where every doctorâs office has a different way of storing data. There are probably 1,000 different electronic health record vendors out there across hospitals and the SMB retail space, and no one generally has more than 3 to 5% market share. So if youâre a developer looking to build any healthcare-data-integrated product, for consumers or B2B, you either build a very niche one, meaning you only serve the orthodontist or only the dermatologist and thatâs it.
Turner Novak:
Yep.
Alamin Uddin:
Or you donât build anything thatâs truly healthcare-data-integrated, because then youâd have to go spend nine months to years looking for partnership agreements with each and every one of these legacy vendors. And even if they do give you a partnership agreement, itâs, âHey, give us 30% of your revenue, or 50% of your revenue,â and you can only have access to a very limited 25% of the data you actually need to build a good product. And you can only read data every 24 hours, and you canât write data. So itâs very limited either way.
Turner Novak:
Itâs terrible. Why would you even say yes to that? I guess you probably have to, thereâs no other options.
Alamin Uddin:
Even those niche players I referenced, that is what they end up doing. By default they already operate within a niche space, dentistry or orthodontics or dermatology, and even within that niche space, those are the types of agreements theyâre signing.
Thatâs why, generally, there arenât amazing products being developed. A lot of healthcare developers get offended when I say there arenât great products in healthcare, but by the standard of your general consumer apps, ChatGPT or iMessage, there arenât a ton of great tools being built. Not because there arenât great PMs or great engineers or great builders in healthcare. Itâs literally just not possible, thereâs just high friction to it.
Turner Novak:
And itâs because of this inability to access the data.
Alamin Uddin:
Exactly.
Alamin Uddin:
Maybe to make it more practical, let me put it another way. If youâre a developer building an app, think of the number of specialties there are in healthcare. There are dermatologists. For me, I have a primary care office, I have glasses so an optometrist, I go to a dental office. There are cardiologists, neurologists. And for each of these specialties, there are specialized electronic health record systems built only for them.
As an example, if youâre a dental office, youâre using Dentrix or Eaglesoft. But if youâre a dermatologist, youâre not using Dentrix or Eaglesoft, youâre using a system called ModMed, which is cloud-based, while Dentrix and Eaglesoft are on-prem. And if youâre an optometrist, youâre not using any of these, youâre using some optometrist-specific system. So the sheer fact of specialization means there are unique vendors for each of them.
On top of that, each of these vendors, even though thereâs federal regulation that kind of mandates it, are also not open systems. So if youâre a developer, thatâs the level of integration effort you have to go through to build a healthcare-data-integrated product for practices, and itâs really high friction. If there are about 1,000 different systems, and you want mass adoption, you probably have to write 100 or so integrations to have a large enough TAM. And even if you do, before NexHealth, anywhere from 30 to 50% of your revenue would go to these companies, so you just donât even have a viable business.
Turner Novak:
What about Epic? I hear people complain about Epic. Theyâre the biggest one in the space, right? They only have a couple percentage points of market share?
Alamin Uddin:
This is a really common thing I get asked about, âHey, what about Epic?â Epic is the famous name. I think they do about $8 billion or so in revenue.
Turner Novak:
They only do $8 billion in revenue? Everyone in the world hates them, and they only do $8 billion in revenue?
Alamin Uddin:
It could be more now, honestly, my figure may be outdated, you can look it up. But a couple years ago it was $8 billion at least.
Turner Novak:
Thatâs crazy. Iâve heard people say, âEpic is the reason healthcare is so bad,â in one sentence, and Iâm like, âAnd they only do $8 billion?â Iâd assume theyâre a trillion-dollar company if theyâre truly destroying the fabric of American society. Come on.
Alamin Uddin:
Thatâs what I mean. They have a lot of brand awareness, and thereâs a perception that they own a majority of the market. What is true is that they do own a majority of the hospital systems.
Turner Novak:
So how is that different than the independents?
Alamin Uddin:
If youâre an average consumer in the US, youâre more likely to go visit a local doctor for your primary care needs, your ambulatory care needs.
Turner Novak:
Versus a full-on hospital.
Alamin Uddin:
Exactly. A full-on hospital would be, think of New York Presbyterian, which is a group of eight to 12 or so actual hospitals throughout the city. Thatâs a hospital system, and Iâm pretty sure they use Epic. Epic owns that market, the majority of it at least. But when it comes to where most consumers in the US actually get their day-to-day care, itâs extremely fragmented.
Turner Novak:
Hmm.
Alamin Uddin:
Epic is not even a player there. Epic is a really interesting topic, though, in terms of why theyâre so popular, or considered the evil bad guys in the industry. A very quick summary is this. Judy Faulkner, who I admire a lot, in addition to thinking sheâs probably an evil genius from my point of view, essentially, in the Obama administration when the Affordable Care Act was passed, there was some healthcare-tech-related legislation along with it.
Turner Novak:
Yep.
Alamin Uddin:
The government was incentivizing practices to digitize their health records.
Turner Novak:
Hmm.
Alamin Uddin:
And the incentive was basically, âWe will give you $40,000 a year if you digitize.â Do you remember, way back in the day, when you would go to a doctorâs office and thereâd be files of paper?
Turner Novak:
Yeah. Sometimes I still have that.
Alamin Uddin:
Right.
Turner Novak:
Actually, thatâs not true. I feel like I havenât run into that in a while.
Alamin Uddin:
We still have customers that are like that sometimes, though itâs very rare now. But at that time, that was the problem to solve. The governmentâs incentive was, âHey, weâll give you $40,000 a year if you digitize your records.â But you have to digitize, and you have to meet certain requirements to qualify for that incentive. And that requirement just happens to be Epicâs spec. In other words, Epic gained mass adoption, especially at the hospital level, by getting the federal government to pay their customers to adopt their software.
Turner Novak:
Wow. Thatâs insane. What a go-to-market. The government does it for you.
Alamin Uddin:
Thatâs what I mean. I admire her evil genius.
Turner Novak:
Geez.
Alamin Uddin:
She pulled it off. Honestly, she really is the Bill Gates of our industry. But at the same time, the hold that Epic has on the hospital ecosystem is not healthy. And I think itâs only a matter of time until this all changes. But thatâs part of the reason Epic is so famous.
Turner Novak:
It might be interesting for people, some might know of you, might know of NexHealth. For someone whoâs literally never heard of it before, first time encountering you and the company, how do you describe it? What do you guys do, and how do you fit into everything we were just talking about?
Alamin Uddin:
The very simple explanation is that weâre the platform for practices and the developers that build for them. What does it actually mean? Today, in our space, we have the largest developer ecosystem, where these developers use us. Honestly, I donât want to say the exact number yet, for competitive reasons.
Turner Novak:
Got it. But a lot. Are you guys the biggest?
Alamin Uddin:
By far. There are a bunch of other reasons we can talk about offline, and we should talk about Epic as an example. But we are the largest by far. The use cases are basically this. You take a company like Align Tech, the makers of Invisalign.
Turner Novak:
Theyâre pretty big, arenât they?
Alamin Uddin:
Theyâre massive. Theyâre out of San Jose. Their use case is, if you go to invisalign.com, you should be able to book with a provider that accepts Invisalign cases. Theyâre using our infrastructure to connect their providers who are trained for Invisalign to consumers. Thatâs one use case.
The other use case, one of the most popular ones these days, is that a lot of the YC startups are built on top of our API to run an AI phone agent. More often the use case is you call your practice after hours, the AI agent picks it up, and you may want to reschedule or cancel your appointment. The AI agent will ask, âHey, whatâs your name, date of birth?â, and once you give it, it needs to know what time your appointment is and then actually action that cancellation or reschedule as an agent. Otherwise itâs useless. And all of that is built on our API and our infrastructure.
The other use case is that we do have a product built directly for these practices. If youâre a practice, you can buy NexHealth, and we will digitize every single step of the patient journey. The key value prop is that if you think of our competitors that came before us, the software was built to help you manage a set of tasks. If youâre an office, to operationally manage a set of tasks.
Turner Novak:
So if youâre the reception woman or the billing person that sits in the front and is always on a computer clicking buttons, with a stack of papers, inputting data.
Alamin Uddin:
Yep. The core differentiator for NexHealth, and itâs only possible because of the infrastructure we focused on building first, is that we donât help you manage the task, we do the tasks for you, literally. And this has been the case for us even before AI. So this isnât because of AI, itâs because of the infrastructure that we own.
What it means is that youâre a practice, you buy NexHealth, and all you really have to do is configure it, set up some rules, turn it on, and every single step in the patient journey gets automated. You donât even have to log into the product, except maybe if you want to respond to patient messages. If a patient responds to a text message from NexHealth, a staff member has to respond back, and that is what we optimize for.
Unlike most companies where the objective is session time, where you want your end users to spend as much time in the software as possible, we optimize for the opposite. We think if our end users are spending a ton of time in our software, weâre failing somewhere, because the core value prop is that we do the work for you so you donât even have to log in.
Turner Novak:
Before you explained this, you said you canât really access any of this data typically. How are you then doing this where you can? Maybe itâs worth talking about some of the first instances of doing that. How did you initially start to build the platform, going back to the very beginning, maybe even before you actually started?
Alamin Uddin:
Happy to explain the full vision here. Our mission statement is accelerating innovation in healthcare. And like I shared earlier, this isnât the first time anyoneâs tried to tackle this problem. Itâs been a problem since the â90s.
Turner Novak:
So people have tried?
Alamin Uddin:
Many, many. Many startups, and even the federal government. The whole reason the federal government cares about data interoperability and digitization of records is that it genuinely wants more innovation in healthcare. And it wants that because healthcare is what, 25% of our GDP at this point?
Turner Novak:
Itâs a huge chunk of the budget. If you go whereâs the money going, Medicare, Medicaid, Social Security, a lot of that is related to healthcare in some aspect.
Alamin Uddin:
Yep. So when I was 22 or so, my co-founder and I were ambitious enough to tackle this problem, and we thought we could. The reason we thought we could is this. Generally, whenever any company has tried to tackle this problem, they take one of two approaches. One is theyâre built specifically for hospitals. And if you build specifically for hospitals, which seems like the large, attractive enterprise market for a startup.
Turner Novak:
Those are pretty big deals, right?
Alamin Uddin:
Yeah, VCs will fund you and so on. Only to realize, after a certain point, that Epic will just kill you.
Turner Novak:
Really? Epic just doesnât allow you to compete? Doesnât allow you to get a foothold?
Alamin Uddin:
Exactly. And by the way, Microsoft tried to solve the same problem weâre solving right now, and they also ran into friction with Epic and Cerner. Cerner is basically the second-largest hospital-based electronic health record system.
Turner Novak:
How are they able to do it if the government wants to solve this? Couldnât they be like, âGuys, cut it out, you canât do this anymoreâ?
Alamin Uddin:
The government tries. Itâs twofold. A lot of the regulation actually doesnât have precedent. You can pass regulation, but you have to enforce it first, and that swings back and forth depending on the administration in power. The second thing is, for the government to enforce it, or for a court case to go through, itâs easier if thereâs precedent of court cases as well, in terms of how you interpret the law.
A lot of the regulation is so new that that hasnât happened enough times to be clear-cut. So what Epic does is they say, âHey, we think regulation says XYZ.â There was a famous battle between Microsoft, Google, and Epic, I think five years ago or so, where Epic was like, âWe think regulation means X,â and Google and Microsoftâs argument was, âWe think it means Y.â And itâs like, who wins? Because thereâs no precedent of court cases.
Turner Novak:
And nothingâs been settled in case law yet.
Alamin Uddin:
Exactly. So if youâre Epic, doing $8 to $10 billion in revenue every year, cash flow positive, you can just take a startup and have them mired in court cases forever, and youâre done.
Turner Novak:
So thatâs one approach.
Alamin Uddin:
One approach to market: hospital systems. And that hasnât generally worked. The other approach is SMBs, which is what weâre targeting.
Turner Novak:
Yep.
Alamin Uddin:
The SMB space is attractive because, unlike the hospital market, the SMB market for EHR systems is extremely fragmented. Why is that attractive for a startup? If youâre looking to build what I call a Plaid for healthcare, the value prop is stronger if you can say, âHey, integrate with us and weâll give you access to 50 or 100 other integrations,â versus Epic or Cerner, which is just two systems.
Turner Novak:
Yep.
Alamin Uddin:
A lot of companies do try to go after the SMB market from that perspective. Then they immediately run into the blocker of either that practice using an on-prem system with no API, or, a lot of the market does use cloud systems these days, Iâd say 40 to 50%, especially on the medical side. But even if itâs cloud-based, the API is extremely limited. You canât just self-serve sign up like most other SaaS vendors. You have to form a real partnership: have conversations, negotiate, get the rates, figure out what endpoints you can access.
So for startups, that 12-to-24-month cycle is just too long. Thatâs one thing. The second thing is, even if you put the partnership, the restrictions, and the incumbents wanting to block data all aside, even if all of that was solved, you still have to write 20, 50, 100 integrations to have scale as a product and a company. All of this to say, those are the set of challenges. But if you can solve it, youâre massively valuable to the ecosystem.
Turner Novak:
Mm.
Alamin Uddin:
I want to distinguish two things here. SMB is more attractive because itâs a harder problem to solve, and if you can solve it, youâre just more valuable to your customers. What ends up happening on the SMB side, though, is most companies before us tried to take the approach of, âHey, weâll partner and aggregate all of these relationships.â They end up building a product where maybe itâs not as good as being able to read and write anything, the way you can with Plaid, where you can read bank data and write back bank data and lots of different types of financial data. Maybe the product isnât as great, but at least they can aggregate all these different systems through partnerships and make it available to developers.
The problem with that approach is that when you sign these agreements with various EHR systems, you end up having to share 30 to 50% of your revenue with them. And if you think about where an API or infrastructure company sits within the software supply chain, your take rate canât be so high that a developer building on top of you has to price themselves out of the market. In other words, they have to price so high because they have to share that revenue with these EHR systems.
Turner Novak:
With the EHR.
Alamin Uddin:
With the EHR. And that in itself literally prices them out of the market.
Turner Novak:
Interesting. So itâs just too expensive for any developers to use.
Alamin Uddin:
Exactly.
Turner Novak:
So when you started NexHealth, I think you were working the first inklings of this. It was like a summer job at a clinic. What were you doing?
Alamin Uddin:
At that time I was exploring maybe going to medical school, or becoming a software engineer, one of two paths.
Turner Novak:
Because youâd learned to code growing up.
Alamin Uddin:
Yeah, in high school, watching YouTube videos. Man, with GenAI now.
Turner Novak:
So much easier than YouTube. The kids these days.
Alamin Uddin:
Itâs crazy if you think about it now. Anyways, I was exploring medical school or something else. One of the things you have to do if youâre looking to get into medical school is show you have some real exposure to the field of medicine, what itâs like to be a doctor. Most students volunteer at a hospital or something. When I was in school, I also needed a job, so I was crafty enough to get a part-time job with a professor at my school who also practices part-time.
Turner Novak:
Okay.
Alamin Uddin:
He was about to retire, so he gave me a lot of freedom, and I was basically the receptionist over the weekend. One of the great things about that job is that you get exposure, because youâre the router of information. Thatâs your job.
Turner Novak:
Every single patient that comes in, you handle it.
Alamin Uddin:
Exactly. And then you coordinate with the labs, the pharmacies, maybe other specialists. So you get exposure to everything. At that time, what I was really trying to do was, âHey, why canât I just automate some of this repetitive communication I have to do?â Like, âDear patient, you have an appointment coming up at XYZ time.â This is 2014, â15. I was like, âOkay, I can probably just build this myself.â I reached out to eClinicalWorks, one of the electronic health record vendors my doctor was using.
Turner Novak:
Mm-hmm.
Alamin Uddin:
Their response was basically, âHey, we donât have an API, but we can give you one for $40,000.â And I was like, âWell, first of all,â Iâm 21 at that time. My insight was, as a college student, if I canât do this, if I canât just hack together something, what about everyone else in the space? One thing led to another and I realized over time that this is a massive problem in our space. And that fit into my motivation for, âAll right, what do I want to do with my life?â, which was a real question for me.
Turner Novak:
Yeah.
Alamin Uddin:
Ultimately it was about impact for me. Part of the reason I was exploring medical school is that whatever I do with my time has to be impactful, from my point of view.
Turner Novak:
Yep.
Alamin Uddin:
Being a doctor is great. However, I think having impact at scale is better. And the best way to have impact at scale, at least today, more pragmatically and something you can control more than other areas, is if you can build something useful for the world, put go-to-market behind it, make sure itâs in the hands of everyone, and really change the world. The iPhone is a great example, or ChatGPT is another. So that exposure to that office, plus the interest and the ambition around impact at scale, led me and my co-founder Waleed to get together and try to build this API for ourselves at first, and then hopefully the whole world.
Turner Novak:
The first thing you built was just some Twilio strings and chains, essentially.
Alamin Uddin:
Exactly. And the funny thing is, I was bashing partnerships, that you have to go through all this. But that is the approach we took ourselves, with eClinicalWorks and a few others. Hereâs a crazy experience. We still donât really have product-market fit, but we have something. Thereâs a large customer that was interested and would give it a try, the first one ever, this is I think 2018. Iâm at the office of that customer, and theyâre like, âWell, we have to call eClinicalWorks to give you access to that data.â So they call, and Iâm live on the phone. Some partnership person from eClinicalWorks finally picks up, and they say, âNo, donât work with NexHealth, we have this solution ourselves.â Live on the call.
Turner Novak:
They didnât know you were there, or they did?
Alamin Uddin:
No, they knew I was there. I had email chains going with them on threads, and we have something signed with them at that time as well. And they knew we were a startup too. They were like, âYeah, just donât use these guys. Letâs set up another call, we have a solution for you.â So thatâs when we realized the partnership approach is not going to work either.
Turner Novak:
Did you guys stay in college? Did you drop out and work on this full-time? How did those next year or two go? It sounds like 2015 through 2018, thatâs quite a bit of time.
Alamin Uddin:
In 2015 I was a sophomore in school. At that time, honestly, this was a hobby more than anything. Weâre still in school. My co-founder and I really got serious about this our last semester at school.
Turner Novak:
Okay.
Alamin Uddin:
What got us really serious, honestly, was when I saw, one, how big the problem was.
Alamin Uddin:
I was naive enough at that time that it was more exciting, if anything. âOh wow, this is such a massive problem. Healthcare is such a large space. If we can solve this, itâll have massive impact.â Not knowing how hard it would be. And then we got some validation, in the sense that every time I would talk to a practice and see the pain myself, this was a really valuable problem to solve. The key question for us at that time was, okay, what do we build?
Turner Novak:
Mm.
Alamin Uddin:
Given we couldnât even build the Plaid, the API product, by partnering with any of these companies, the real question was, how do we go about this? What do we build, and what should be our first product? Historically, before us, a ton of companies tried an API approach where they tried to build a platform for healthcare, and it didnât work.
To fast-forward, the strategy we landed on is this. These incumbents are not open systems. Theyâre the gatekeepers to innovation, and theyâre just not incentivized. Everything comes down to incentives. Theyâre not incentivized to be an open platform and give you access. What are they incentivized by? If youâre Epic, if youâre any of these companies, youâre incentivized to keep your customers paying you and generally happy with you. Some donât even care about that, honestly, but generally.
So thinking from first principles, our job is to make sure that those same practices that pay these vendors love us, want to use us, and give the incumbents no choice but to either tolerate us or partner with us. But the way you gain that outcome is this. From first principles, what do these incumbents, Epic and Cerner, or Dentrix and Eaglesoft in dental, have that we donât have? They have really direct relationships with these practices. So in terms of what we build, letâs build something genuinely useful for these practices, that either makes them money or saves them money. That has to be the incentive for the practice owners.
As long as these practice owners value our product, they become dependent on it, where it has to be so good that they canât live without it. And as long as we can quickly gain enough mutual customers with each of these systems, the incumbents will have no choice but to partner with us or support us. Hereâs an example. You look at a company like Henry Schein, one of the larger ones. Theyâre like the Epic of dentistry. I can go into what that means exactly.
Turner Novak:
Iâve heard of them before.
Alamin Uddin:
We have something like 10,000 mutual practices with them. We donât have a partnership with them or anything like that. We bypass them, and we go directly to the practices and the database, or if itâs cloud-based, directly to the browser. Over time, the evolution of our relationship with them has been this. Initially they really tried to block us, because we had maybe 100 customers with them. Then in 2023 they went as far as sending a mass email to their entire customer base and showing an in-product pop-up to half the dental space, telling them not to use NexHealth.
Turner Novak:
A pop-up in their product, youâre saying?
Alamin Uddin:
Yeah. And an email. At that time it was terrifying. I was like, âWell, I guess itâs some brand awareness,â because at that time no one even knew who we were. And then over time itâs evolved to now neutral, to, âHey, maybe we should partner.â So over time that problem, where they were very hostile and really just trying to kill us, has been solved.
The reason itâs been solved is because there are 10,000 practices out there that use NexHealth for their scheduling, their payments, their paperwork, their patient communication. And for Henry Schein, the question is, do we want to turn off 10,000 practices that really rely on NexHealth for their day-to-day operations? Are those practices going to be happy? Of course not. So the reason we started with our SaaS product first is to gain that incentive alignment one way or another. Of course Henry Schein isnât happy about it, but their incentive is now aligned with ours to a degree, where both of us just want to support our customers. So it probably makes sense for us to at least talk and maybe partner with each other.
Taking that approach of building a SaaS product, making sure it grows as quickly as possible, and then releasing the API product means our SaaS offering lets us actually protect the API product, and make sure no developer that builds on top of our API gets disrupted by these incumbents. Does that make sense?
Turner Novak:
Maybe. So by building the SaaS, you build relationship and trust, and then you build the API on top of that, and that enables what specifically?
Alamin Uddin:
What it enables is this. If youâre an EHR system, for context, maybe Iâll use a Plaid example. Do you remember a couple years ago there used to be headlines like, âCapital One is blocking Plaid,â or, âChase is blocking Plaidâ?
Turner Novak:
I sort of remember this.
Alamin Uddin:
We have a very similar dynamic in our industry, where these incumbents will release some software update that basically tries to block our access to the EHR system.
Turner Novak:
Oh, itâll be disguised as, âOh, itâs just an update.â
Alamin Uddin:
Exactly. âSorry about that.â It used to happen every three months or so. They donât do that anymore, because there are just so many of their customers reliant on our software now.
Turner Novak:
And so that makes the API better, because developers can rely on it.
Alamin Uddin:
Exactly. And connecting it back to why startups before us didnât succeed with the same product: they just didnât figure out a way to align the electronic health record vendors and their incentives to support an open ecosystem. We kind of forced their arm.
Turner Novak:
Made it happen.
Alamin Uddin:
Yeah. So if youâre a Plaid for healthcare, all you are is an API company. And if youâre Epic, you can just be really hostile and push some update that blocks your access completely. That means every app built on top of your API is also disrupted, and you just go out of business. Versus with NexHealth, they can try to do that, but we have direct relationships with these practices. We recover really quickly, and honestly weâve never really had any downtime there either. But more generally, the bigger we get, the more rarely it happens, because you just donât want to disrupt the whole market if youâre Epic.
Turner Novak:
So itâs basically like you have to get the trust of the end customer.
Alamin Uddin:
Mm-hmm.
Turner Novak:
I think the number I saw, maybe itâs bigger today, thereâs something like 89 million people that touch the NexHealth product. Something like 25% of the population.
Alamin Uddin:
Yeah, 89 million consumers.
Turner Novak:
Thatâs insane. Talk about having a big impact on the world. Hopefully the NexHealth product is good, but if so many people are using it, itâs probably pretty good.
Alamin Uddin:
Yeah.
Turner Novak:
Back when you were doing the very first customer, the very first integration, how did you get around this whole âwe canât access this, thereâs a server in the closetâ thing? How did you make that work?
Alamin Uddin:
Oof. One of the things we did, and this will sound like a general framework for how to find product-market fit, putting the integration question aside, but that approach also helped us solve it. My co-founder and I were in New York at that time. Thatâs where we grew up. We decided that all we were going to do, in terms of finding product-market fit, was focus only on one specific zip code, which was Soho at that time in New York. And I actually went into the offices to pitch them our product.
Turner Novak:
And this was all dental?
Alamin Uddin:
Exactly. The reason we chose dental specifically is that medical is way more fragmented than dental, in terms of a vertical specialty thatâs large enough to focus on and have a meaningful business to start as a wedge.
Turner Novak:
So the EHRs were concentrated enough that you didnât have to do that many integrations, and you could still make enough revenue fast enough at enough scale to justify it.
Alamin Uddin:
Exactly. In dentistry, if you integrate with five systems, your TAM is probably around 100,000 practices, versus in medical, if you integrate with five systems, your TAM is probably 20,000 practices. Itâs that wide. So we chose dentistry, and we thought itâd be a good wedge for us. Itâs highly competitive, though.
Turner Novak:
Competitive in what way?
Alamin Uddin:
There are just a ton of vendors in this space, for similar reasons to us, focused on dentistry. So we would go into the offices, try to sell them, and then say, âOkay, you use Dentrix, show us your server.â At that time we figured out, thinking from first principles, itâs a SQL Server in your office. Why canât we just run an agent that runs a cron job every 15 minutes, some cycle?
Turner Novak:
This is a pre-AI agent, right? This is a recurring task.
Alamin Uddin:
Yeah, thatâs the old definition of an agent. A cron agent on the server. So weâre like, why canât we just do that, as long as the practices consent to it? And a lot of practices would consent to it.
Turner Novak:
Really?
Alamin Uddin:
Generally, yeah. Because they know itâs their data, they own the hardware, they pay for it. And as long as they trusted us, theyâd give us access to their server, and weâd reverse-engineer the whole schema in terms of how to read and write the various patient data, run this cron job on that machine locally. Thatâs honestly the approach we tried. We were just going to go into the office, get access to the server, without an API or official access.
For on-prem, the traditional API method doesnât really exist. But more generally, without going through the actual vendor that makes the software, just go to the local machine, because thereâs a local machine that stores that data, and read and write into that local machine at the doctorâs office. The only reason we even got to that insight is because we would go in person and talk to a ton of practices, either to convince them to use our software or to learn more. One thing led to another, a couple of them let us do it, and thatâs how we scaled to this day.
Turner Novak:
That sounds extremely not scalable. Literally go to offices. So what would you recommend, in terms of knowing this isnât a dead end? In the process of this, were there certain indications you were getting as you went that this strategy could work?
Alamin Uddin:
To be clear, if you donât have a product to begin with, whether youâre doing scalable things or not doesnât matter. Who gives a shit?
Turner Novak:
Itâs still just making the product.
Alamin Uddin:
Yeah. Weâre still just figuring it out. Remember, at that time we didnât even know, because the partnership approach we took with these vendors first just didnât work. They didnât want to partner with us.
Turner Novak:
Okay.
Alamin Uddin:
So itâs like, okay, our product doesnât have any value unless we can read and write into those on-prem local machines. So we donât really even have a product. The approach we took was, weâre just going to canvass the neighborhood, go into these offices, and see if we can run some local agent on that machine thatâll let us run our cloud-based software. So you can go to NexHealth.com, log in, but the data exchange is happening locally, some cron job running at that office thatâs connected to our AWS backend. Thatâs the way to think about it.
In terms of scalability, the reason we would go into the offices and physically, manually configure a bunch of this is that it was needed to gain trust from the offices, showing up in person. And second, thatâs what was needed to go through the discovery process of how to reliably read and write into the systems. But in terms of scalability, we always knew we could productize it. You download Spotify on your MacBook, or Notion, or any software on your MacBook. Very similarly, you can go to nexthealth.com/download and download NexHealth if you use an on-prem, server-based electronic health record system. You just download it, consent to a few things, and within 10, 15 minutes youâre up and running today.
Turner Novak:
So youâre basically mirroring their on-prem server into the cloud. Any time thereâs a change to the server, it eventually gets updated to the cloud, and maybe it got faster over time.
Alamin Uddin:
Thatâs one way of saying it, yeah.
Turner Novak:
I say it that way because Iâve seen that approach with some of these system-of-record companies that are kind of AI-native. You run your traditional cloud system of record, and then you have your AI one, and maybe you donât have to adopt the AI one right away, so they just mirror the cloud-based one. You use the AI one to do some work, but you still have the other one, and itâs easier to transition customers over and get them to try it without full adoption. Mirroring the system of record, I guess, is the way to do it.
Alamin Uddin:
I thought you were saying that as an analogy, not literally. But actually, no.
Turner Novak:
Oh, so you did not do that?
Alamin Uddin:
No, we donât do that.
Turner Novak:
So how did you do it originally?
Alamin Uddin:
Hereâs what I mean. Every cloud-based software company has a Postgres database. What our customers and these practices have in their offices is an equivalent of a Postgres database running locally.
Turner Novak:
Okay.
Alamin Uddin:
And thatâs where all the patient data exists. Our job is to sync the data back and forth. Thatâs all it is.
Turner Novak:
This is literally like a file thatâs just saved on a computer.
Alamin Uddin:
Exactly right. So our job is to figure out, what is the schema of that database in the office? Youâre a patient, you have appointments, contact information, prescriptions, past appointments, family, and so on. Thereâs a whole schema set up for you in that local database. And if weâre going to read that data, but more importantly write that data, we of course have to insert the correct data in the right tables in that Postgres database. They donât actually use Postgres, Iâm just using it as an analogy, I wish they used Postgres. So we have to read and write and sync data between some local server at some office and NexHealth, which is cloud-based software.
Turner Novak:
Yeah.
Alamin Uddin:
Without an API from the vendor that makes that local software. And the way we do that is we run a piece of software on that machine itself that connects to our cloud-based system and syncs the data back and forth, to put it very simply.
Alamin Uddin:
We were young and had a lot of time on our hands. We were really passionate about the problem. Honestly, it was really interesting to us.
Turner Novak:
Yeah.
Alamin Uddin:
These days, with GenAI, software engineering you donât really have to learn anymore. But a couple years ago, software engineers more generally donât even have to deal with on-prem systems, networking, and so on. So to us it was interesting from that perspective as well.
Turner Novak:
Yeah.
Alamin Uddin:
Weâre SOC 2 compliant. We do our annual HIPAA audits as well, so all of it is compliant and secure. And in fact, a ton of these EHR vendors have a ton of security incidents, and to this day I donât think weâve had any that I can think of.
Turner Novak:
Really?
Alamin Uddin:
Yeah. Weâve been able to productize it to that level, encrypt it to that level, and bring that level of authentication and security. And more importantly, make it reliable enough that we have never corrupted a database to this day.
Turner Novak:
Wow. So all this stuff you were doing, had you raised a little bit of money to get started?
Alamin Uddin:
You mean early on?
Turner Novak:
Through this whole process of 7:00 PM, staying up till 1:00 AM, did you raise any money to do this? Were you guys working other jobs to fund this? How did this all go from a capital-stack side?
Alamin Uddin:
Honestly, at that time, my co-founder was selling used computers or something on eBay, and I still had my part-time job while we were doing all of this. So at that time we didnât have any funding.
Turner Novak:
Were you making any money from customers?
Alamin Uddin:
No. Remember, we didnât even have a product to begin with. So maybe let me explain what the product is. The original insight was, we have to build a relationship directly with the end users, meaning the practices that own the hardware and the database and pay these EHR vendors. How do you do that? You have to build a product thatâs revenue-generating. So we built this scheduling tool. An online scheduling tool wasnât new either. Zocdoc and a bunch of competitors were already dominating the market. But where we differentiated is that we can read your real schedule in the office, which runs off that on-prem system. We can read your real patient data, and then we can actually write back as well. The write-back part was the most differentiating part at that time.
What it basically meant is this. Letâs say youâre a dermatology office or a dental office, or a PE roll-up with 10, 15, 20 offices. Oftentimes youâll run some paid advertising to attract patients. You have Google Maps, Yelp, social media, your website. And the business model of every practice ultimately comes down to how full is each room, and is each available slot for each provider filled or not?
Turner Novak:
Because itâs basically a capacity-utilization thing. You pay rent, you pay the doctor, the nurse, etc., and thatâs fixed. You donât just say, âHey, weâre just not going to pay.â You need to have the rooms full of patients.
Alamin Uddin:
Exactly. And in the tech world, we donât really think about these things, because itâs software, you can scale hypothetically infinitely. For a practice, thereâs a ceiling to their growth, and after that it comes down to utilization optimization, and then efficiency in terms of how you optimize it. Especially in the larger cities, this happens often, where they spend some money on marketing and social media. So we built a scheduling product where the value prop was, we can read real data. If thereâs a slot open, weâll know it within 30 seconds, and if thereâs a patient that wants to book, weâll collect all their information and write it back to the calendar exactly the way one of your staff members would. And that logistics, in terms of a product build, is way more complicated than youâd think.
Turner Novak:
Well, this was pre-generative AI too, right? So the schema of how data is transferred, you have to get it one-shot perfected every time.
Alamin Uddin:
Honestly, even the product design. Let me give you an example. Letâs say you go to a dental office. What do you do? You first see a hygienist.
Turner Novak:
Yeah.
Alamin Uddin:
And then you see the dentist. So as software, you have to coordinate those times, and what is their availability.
Turner Novak:
So itâs really two appointments, in a way.
Alamin Uddin:
Yes. And oftentimes itâs in two different rooms. So you also have to take into consideration that this provider may already be with another patient in some other room. And by the time the hygienist finishes with you, that provider may not be available. So itâs harder than youâd think. This is the write-back part. Itâs not only the data access, itâs also the logic. Anyway, that part we figured out, and thatâs the part we went to market with.
The value prop was, if your practice slot utilization is really what youâre optimizing for, one, you can convert new patients across the internet, just put NexHealth there. As well as, we can reach out to your existing patients. If youâre a patient and you want to come in sooner for some reason, and youâre on a wait list, and another patient cancels today at 2:00 PM, NexHealth will automatically text you.
Turner Novak:
Kind of fill up the wait list.
Alamin Uddin:
Exactly, on a first-come, first-served basis. And that was incredibly popular with the practices. Versus Zocdoc, which is the more dominant scheduling tool in our space. Iâm not sure if youâre familiar with them.
Turner Novak:
Vaguely. I know of it. Itâs a yellow logo, right? Yellow and black.
Alamin Uddin:
Exactly. Theyâre more like a Yelp kind of marketplace.
Turner Novak:
Yeah, I was going to say itâs kind of like an SEO marketplace, lead-gen type thing.
Alamin Uddin:
Exactly. So the way to think about our product is that we augment your staff. The patient doesnât have to call, they donât have to do anything manually. And once we gain adoption, itâs really hard to rip out. If you rip it out, then some staff member now has to spend something like 30 to 45 minutes coordinating those calls from patients, rescheduling appointments, filling slots. So what ended up happening is, as practices started using NexHealth and loved it, we started growing fast enough that by the time any of these EHR systems even noticed NexHealth to begin with.
Alamin Uddin:
There were enough practices that loved us enough and advocated for us enough that these EHR systems just couldnât shut us off.
Turner Novak:
Interesting. And people were paying for it at that point, right? Customers were?
Alamin Uddin:
Yep.
Turner Novak:
I think I saw a number, you raised $391,000. Is that the number? That was the very first amount of capital you raised. Thatâs an odd number. How did that come about? Was it just scraped together, like one person gave you $6k, and thatâs how?
Alamin Uddin:
Man, this is bringing back memories. I think I buried that round in my memory. It was that hard.
Turner Novak:
Thatâs always the hardest round to raise. People will say the IPO is the hardest, or whatever round.
Alamin Uddin:
Thatâs a champagne problem, the IPO being the hardest. You have to think about this. We didnât actually have any connections to Silicon Valley or the Bay Area at all, any VCs. My co-founder and I grew up with immigrant parents, so honestly we donât really have any connections to money to begin with. So the way we raised that initial capital was two sources. One was actually a couple of our professors at school. And then a couple of our customers on top, and thatâs how we scraped together $390,000. We still have two of those customers on our cap table, still using us, love us. So honestly it was our customers, and then people we knew, professors, a couple of friends. Thatâs how we scraped together $390k.
Turner Novak:
Wow.
Alamin Uddin:
And then a year later, we ran out of cash, so it was terrible from that side.
Turner Novak:
Really? How did you run out of cash?
Alamin Uddin:
When we raised that capital, we raised it off of early customers, maybe the first 10 to 20, a few of them loving our product and seeing potential in us. And the second part was people we knew, professors and so on, also seeing potential in us. But we didnât really have any amount of revenue that could actually sustain a business to begin with. If you take 20 customers, thatâs probably a couple thousand of MRR a month for us.
Turner Novak:
So you were still burning money?
Alamin Uddin:
Exactly. It was me and my co-founder, and we hired two or so engineers. And we talked about earlier what it took for us to figure out how to read and write to these on-prem systems, and you were surprised yourself, how do you even scale that?
Turner Novak:
So you were figuring it out with the $390,000.
Alamin Uddin:
Yes. And it took us a year-plus to get to some level of stability with it.
Turner Novak:
Okay.
Alamin Uddin:
Where even if we would sign customers, they would actually end up churning.
Turner Novak:
Man, that sucks. Whyâd they sign up and then churn? What was the issue?
Alamin Uddin:
They would sign up because of the potential of what it could do. It was something new to them, so they were early adopters. Theyâd give us a try for a couple of months, and then eventually, if it doesnât work, it doesnât work. It wasnât stable enough. The thing about our value prop is that if youâre going to have your ads running and youâre spending money on it online, and NexHealth is the CTA for it, where as a patient you see, âHey, we do Invisalign,â and you click, you fill in your information, it has to be reliable. If the value prop is that your staff doesnât have to monitor that channel, and it just writes back to their on-prem system, it has to do that. Otherwise youâre losing money as a practice, and losing patients.
Turner Novak:
Because youâre not generating any revenue from those ads.
Alamin Uddin:
In fact, youâre probably causing them to lose money. When I say the stability of the integration: if youâre a patient, you click on some Facebook ad, fill out a form on NexHealth, and you can see the real-time availabilities from that practice. You pick a time, tomorrow at 2:00 PM, and fill in your information. We take that information and have to sync it back to their on-prem system. Oftentimes, writing it back just wouldnât work for one reason or another, and that patient and their information would just not go through. So the patient would show up and theyâre like, âWhat are you talking about? We donât have your appointment.â It was that sort of problem.
Turner Novak:
Oh, wow. So itâs like, we need to get rid of this, this thing is causing chaos in the office.
Alamin Uddin:
Exactly. All this to say, the real problem, honestly, was that we werenât focused on making any one system and its integration work first. We were trying to make five of them work at the same time. So after nine months of trying, we realized, you know what, we just have to focus on one system and make that great. And as we started making that great, customers would stay with us, sign with us, they were happy, they were referring customers. But we ran out of money. So it came at the nick of time.
Weâve honestly always had product-market fit, because our mission statement is accelerating innovation in healthcare, which means by default there arenât a ton of competitors to begin with. As long as we can make the data integration work, and it sounds so basic, whenever I talk to people in tech theyâre like, âOkay, whatâs the big deal?â, but for our space it doesnât even exist today, itâs all manual. So we knew that as long as we could make that work, any product we build on top would be valuable to customers. And as soon as that clicked for that one system, we started growing really quickly. But a year in, our cash balance was basically $4,000.
Turner Novak:
Like you were going to die in a week or two.
Alamin Uddin:
Well, at the end of the month. It was that sort of situation.
Turner Novak:
So how did that month go?
Alamin Uddin:
Thinking back, sometimes people ask me, âHey, whatâs the closest NexHealth came to death?â Throughout our history, that was the only one where it truly was a consideration for me and my co-founder of, âHey, weâre still young, maybe we just get a job.â
Turner Novak:
Yeah. You had a bunch of credit card debt too, right? I read on the website you had essentially maxed out your Amex card. Your credit score was 500-something?
Alamin Uddin:
Yeah, 500, like, 30-something.
Turner Novak:
And you got a court order from Amex or something? Basically, you must pay this bill.
Alamin Uddin:
Yes.
Turner Novak:
So you really had negative money. You had $4,000 in the bank, but you had way more credit card debt than that.
Alamin Uddin:
Yeah. Basically what was happening is that $4,000 stayed in the bank because we had to pay dues and so on. It was just enough to pay the bills.
Turner Novak:
Yep.
Alamin Uddin:
But everything else, Iâd be like, âAll right, let me just max out this credit card.â At that time, honestly, I didnât really think about the personal risk. From maybe just a personal belief that everything would be fine ultimately. And then, of course, it worked out for us. To this day, by the way, it still kind of impacts my credit score. Yeah, fucking Amex.
Turner Novak:
So I think you guys essentially landed a customer that really carried you.
Alamin Uddin:
Basically the reason I had to max out my credit card is this. At that point we had two engineers, and Iâd hired this kid out of college as an SDR as well. We couldnât pay them, and of course we couldnât pay ourselves.
Turner Novak:
Yeah.
Alamin Uddin:
Out of those two engineers, one of them left completely. The other, by the way, is still with us.
Turner Novak:
Really?
Alamin Uddin:
He was okay to delay his payroll for a couple of weeks. It ended up being three months. And then everything else was just through my personal credit card. Throughout that whole period, two things were happening. One is my co-founder and I were genuinely having conversations of, âMan, this is really hard.â Think about it from our point of view. We first tried the partnership route, spent a year on that, and it didnât work. Then we tried going to the database directly, and at that time I was like, is that even going to work? Because it took us a year to churn through customers and figure it out.
Turner Novak:
Yep.
Alamin Uddin:
This is hard. So maybe we should go back to school. And I remember, this is frankly one of my biggest life lessons. At that time, I donât know why, but Iâve always been, itâs kind of cliche, but organically more of a first-principles thinker. So I was just asking myself, okay, why do companies or ventures fail in general?
Turner Novak:
Mm.
Alamin Uddin:
What is the definition of that failure? At what point can you say it failed or it did not fail? Is it when you run out of money? Is it when you donât have a product? How do you define that? And I realized itâs actually a bunch of people just admitting they failed, or giving up. And I was like, well, as long as we donât give up.
Turner Novak:
You just donât have to give up if you donât want to.
Alamin Uddin:
Yeah. We donât have to give up. Itâs that simple. It wasnât that simple at the time, of course, but if thatâs the definition. And at that same time, this customer, whoâs still with us by the way and is massive now, but at that time was only 12 locations in Detroit, I was working on a deal myself, directly negotiating with their CEO. This was I think February of 2018. Heâs like, âOkay, we love the product. How much is it? I hear thereâs no contract.â Thatâs what Iâd pitched him on, that itâs month to month, they can cancel anytime. And I was like, âYeah, itâs a new year, we changed our policy, you have to pay up front now.â
Iâm not going to name the CEO, but he knows all of this at this point. Honestly, Iâm grateful to that customer and that CEO. They saw the value enough that they were like, âOkay, sure, weâll pay you $36,000 up front.â And that kept the company afloat long enough that Adam Lin, whoâs one of our actual first VC checks in the Bay Area, gave us enough capital that we were able to sustain ourselves and find product-market fit.
Turner Novak:
That CEO is probably solely responsible for accelerating innovation in the healthcare industry.
Alamin Uddin:
100%. He recently came to our customer advisory board. And I was like, âDid you know X, Y, Z?â And heâs like, âI did not know that.â
Turner Novak:
You famously didnât really tell your investors this plan, how you were basically going to go from SaaS to the API to building this full platform around it. Why didnât you, or did you try?
Alamin Uddin:
At the very beginning, that friends-and-family round we raised, as I would share our grander plans, basically no one would take us seriously. I understand from their point of view. First of all, youâre just two kids, no experience, and itâs healthcare, where nothing really succeeds. Distribution is hard, and youâre telling me that.
Turner Novak:
You have no customers yet.
Alamin Uddin:
Yeah, and a product that doesnât work either.
Turner Novak:
Oh, yeah. Geez.
Alamin Uddin:
So I donât blame them. But honestly, thinking back, if you just look at our strategy, itâs kind of incredible that it works today. Theoretically, I can see why youâd think itâs not going to work. Weâre going to build a SaaS product, build it to get the customers on our side, so that ultimately we can release this infrastructure product that anyone can use to innovate on top of, and then take both audiences and turn it into a full two-sided platform.
Turner Novak:
Mm.
Alamin Uddin:
Thatâs two businesses in one company, first of all. Second, itâs two different audiences. And third, a lot of things have to go right in just the right sequence for you to even be successful, and itâll take a long time. So if youâre a traditional VC or investor, I can see thereâs an extremely small chance that would work. And I probably wasnât explaining it the right way. So what I decided, the next time we raised properly, which was 2018, was, âHey, weâre just going to build a great fucking SaaS business first.â
Turner Novak:
Software for dentists.
Alamin Uddin:
Exactly. Make sure it grows fast enough, customers love us, the retention is high enough, itâs efficient. That in itself gives us enough credibility to then tell the bigger story. And that ended up working.
Turner Novak:
Interesting. So this is 2018. I think you raised a couple million dollars. How did those next two years go before COVID?
Alamin Uddin:
In 2018 we raised a couple million, and that was purely off of 100 or so customers. Really highly differentiated product. Customers were honestly just obsessed with our product.
Turner Novak:
Thatâs got to feel really good, going from everyone churning a year before.
Alamin Uddin:
Exactly. Thatâs what Christophe at Point Nine saw early on about NexHealth. At a minimum, this is a product that customers love. They knew about the grander vision as well, but at a minimum, the downside is weâre a great SaaS business. And then from there, our goal was twofold. One is, remember, as of now we only support one integration, which at the time is too small, so write more integrations to begin with.
Turner Novak:
So you really still just had one single integration. Did you have to find dentists that had that EHR?
Alamin Uddin:
Yeah. So what we do, and to this day we do this, is we go to market and say, âHey, weâre NexHealth, we have this product.â And remember, this only works if your product is actually differentiated. You go to a practice, you donât actually support their practice management or electronic health record system yet, but you say, âHey, we have this phenomenal product, and if you become a development partner,â a dev partner as we call it, âweâll give you a certain number of months for free and some incentives. But ultimately youâll have this great product, and youâll get to partner with us in building this integration.â And thatâs how weâve done all of our integrations.
Turner Novak:
Interesting. So you basically say, âWeâll give you access to this for free for a while if you...â
Alamin Uddin:
Honestly, three months, generally.
Turner Novak:
And then thatâs how you built the new integrations?
Alamin Uddin:
Yep.
Turner Novak:
How many are you at today? How many integrations does NexHealth have across everything?
Alamin Uddin:
If you consider everything, I think 21 or so.
Turner Novak:
21 different EHRs?
Alamin Uddin:
Yeah. Publicly, in terms of what developers can actually use, itâs 15 or so, and there are a bunch of other ones. There are different tiers we have, some public, some not. But in total, 21.
Turner Novak:
Whatâs the business? How many different business lines do you have, different products? You have something called Synchronizer. Itâs kind of like an API-type platform.
Alamin Uddin:
On the 21 side, earlier I mentioned we have to have 100 or so to get mass coverage. Hereâs the fascinating part I discovered over time myself. One, the market is massive to begin with. Just dentistry itself: Henry Schein, the company I mentioned earlier, is mostly a dental-oriented company, thatâs the vertical they focus on, and they do $10 billion a year. So you can build that size of a business just within dentistry.
Going back to the earlier lesson on focus, we do have a culture of genuine focus and not overextending ourselves, as a result of that early experience. So today we have customers across 19 different specialties, with 21 integrations. You can think of dermatology systems, primary care, dental, across the board. We support a wide set of specialties outside of dentistry today.
Turner Novak:
Mm.
Alamin Uddin:
And as we speak today, we have three lines of business. The first was of course the SaaS product we built, where we charge a monthly fee to practices. Around late â24, I started to see signals in the market that weâve de-risked the existential risk from the EHR vendors significantly enough that itâs time to open up our API, our technology, to third-party developers. Anybody can now go on our website, click sign up, and start building healthcare-integrated products that read and write to 15 different systems, giving you a TAM of 400,000 to 500,000 practices. So weâre very targeted there.
And that product, within 18 months, in terms of pure number of practices it supports, has gotten to the same scale as the SaaS product. For the SaaS product, it took us five years to get there. For the API product, it took us 18 months. And today, on a pure scale basis, the API product is larger than the SaaS business. And then in late â25, we released a payments product for practices specifically.
Turner Novak:
To accept payments from customers?
Alamin Uddin:
Exactly. Have you ever gotten the statements in the mail from your doctorâs office?
Turner Novak:
I lose them and hate them. Actually, what I get most often is Iâll get the collections bill, and Iâll just be like, âI didnât even see the bill. I didnât even know I owed you money. Why are you sending me bills in the mail? Iâm a millennial, I donât read that. Just shoot me an email or a text and Iâll pay it.â
Alamin Uddin:
Honestly, the world of healthcare practice is so insane that we can talk about this all day. Hereâs how payment collection works at an office today. The dominant players in our space when it comes to payment terminals and processing are your typical ones, Square and a bunch of others.
Hereâs an anecdotal example to exemplify how big the problem is. We have this customer a couple of minutes from here. About a year ago, they were one of the early adopters of the payments product. They use Square. I went in just to talk to the end users, and I was like, âHow do you like it?â They love the product. âWhy do you love it so much? What did you used to do?â
Whenever I go to our customers, I also have a real visit, so I got a real cleaning too. I had to pay like $50. Sheâs like, âYou know what, Iâll pull up our old system and process the money through that old Square and show you what we have to do.â So I had this appointment at 1:00 PM. She reactivated the Square terminal and swiped it. âOkay, so you paid this $50.â At that point it was like 1:45.
Turner Novak:
Yep.
Alamin Uddin:
Sheâs like, âOkay, so at the end of the day, what Iâm going to have to do is take the receipt, see that itâs $50 around 1:45, make a guess that, âHey, Alâs appointment was at 1:00 PM, so this $50 payment must be from Al,â and then manually type it into the electronic health record system.â
Turner Novak:
Oh, because itâs not synced.
Alamin Uddin:
Exactly. And not just not synced, they have to guess who paid that bill. So why didnât you get your bill in the mail, or why did you just get a collection? Itâs because often they donât even know.
Turner Novak:
Wow. So even when you do pay right at the point of sale with a Square, itâs not connected to the EHR. So are there ever mistakes where your bill doesnât even get paid properly?
Alamin Uddin:
A ton. At any given time, an average SMB practice will carry like $45,000 every month in AR, and then at the end of the year they write off like a few hundred thousand to bad debt.
Turner Novak:
Really?
Alamin Uddin:
Yeah. And here are the intricacies of the whole system. That cash payment I just made, along with everything else, the reason it has to be written back to that electronic health record system is that the EHR has a component that connects back to, to simplify it, Aetna or whatever payer they use.
Turner Novak:
All the insurance information.
Alamin Uddin:
Yeah. And that practice has to submit a claim to get reimbursed on the rest of the visit amount. Whatever you owe, say $500, you paid $50, so now the insurance carrier maybe has to reimburse that practice $450. All the data has to be perfect for you to even get the reimbursement. Any error, and of course, if youâre an insurance carrier, youâre kind of incentivized to not approve a lot of these.
Turner Novak:
Or you haggle it down too, right?
Alamin Uddin:
Exactly. So the pain and level of inefficiency for these practices is that high. What this practice will do today is twofold. One is, thereâs a concept in our space called admin day.
Turner Novak:
Whatâs it called?
Alamin Uddin:
Admin day. Administrative day. The typical thing is every Friday theyâll either not be open at all, or close the office at 12:00 PM. And what do they do the rest of the day? Administrative work, meaning theyâll type data into their electronic health record system, or send out the statements to patients through the mail. Oftentimes theyâll have some external agency that owns a lot of this for them, where they spend a ton on paper and postage.
And this is the power of the Synchronizer API that we have, the core infrastructure we built that sits on top of all these electronic health record systems, reading and writing data. Whatever product you build on top of it, whether youâre NexHealth or an external founding team, you gain instant 10x differentiation by the sheer fact that you can remove hours and hours of manual work, errors, compliance risk, and lost revenue from a practiceâs workflow.
With the NexHealth payment terminal, if Iâm a patient and I come in at 1:00 PM and have to walk out at 1:45, NexHealth knows you have to walk out at 1:45. All the practice has to do is click once, âHey, I want to collect this balance,â and we auto-calculate everything. Thereâs a whole calculation process for how much you owe, your copay, that theyâd otherwise do manually. We auto-calculate all of it, and in near-instant real time show you the balance on the terminal without any other manual work, and we know itâs you. You swipe it, and then we type all the information back to whatâs called a ledger in the EHR system, and then we zero it out as well. So thereâs a whole accounting concept that plays into this. So what used to take hours and hours, like a half day, is now a couple of clicks, literally a couple of seconds.
Turner Novak:
Interesting. So you said the software product took about five years to get to the scale that took the API 18 months. How has payments been going?
Alamin Uddin:
I donât want to share exact numbers, but in the last 12 months it grew 308%.
Turner Novak:
Itâs grown even faster than the API product?
Alamin Uddin:
Yep.
Turner Novak:
Interesting.
Alamin Uddin:
Honestly, I can talk about this product all day, itâs one of my personal joys. The value prop we offer these practices is so massive that generally, if customers hear about it, they donât first believe it.
Turner Novak:
Really?
Alamin Uddin:
Yeah. Theyâre like, âHow do you even do that?â
Turner Novak:
So what does the average doctorâs office look like that hasnât heard about NexHealth yet and thinks, âOh, that just sounds fakeâ? What does their technology stack look like?
Alamin Uddin:
There are two types. After COVID, like we talked about earlier, and especially the last couple of years, the ICP profile has really changed, and I think thatâs also whatâs leading to some of the faster growth weâre seeing. But still, you have to realize these doctorâs offices are so busy, and they do genuinely get bamboozled by salespeople all the time in our space. So thereâs a bit of cynicism as well.
Turner Novak:
Mm.
Alamin Uddin:
However, we have a great brand in the space, and customers generally trust us. At the same time, theyâll be skeptical, like, âHey, I pay an entire agency, I have half an administrative day just to do this work, so how are you guys turning all of that into a couple of seconds?â Thatâs the skepticism. But once we show it to them, our win rate there, on pure sales math, last I checked was 72%, which is wild. No one has a 72% win rate on any product. Honestly, when I saw it I was like, âIs this data entry? Is this a Salesforce hygiene problem? Like an error?â Itâs not. Itâs wild.
Turner Novak:
You mentioned COVID again. So what happened inside the company when COVID hit? Because youâd only raised a couple of million bucks, and itâd been a couple years since you did it.
Alamin Uddin:
Hereâs the chain of events. We raised that couple million back in 2019 or so. And then within 18 months, COVID hit. Right before that, though, we had a term sheet signed for our Series A, right as COVID hit.
Turner Novak:
This is the beginning of COVID, and it wasnât closed yet?
Alamin Uddin:
It wasnât closed yet either. So thatâs how intense that period was.
Turner Novak:
So when did you sign the term sheet?
Alamin Uddin:
We literally signed the term sheet, I think, end of February 2020. And if you recall, it was the first week of March where everything was shutting down.
Turner Novak:
End of February, people were making fun of the sign on a16zâs door saying, âNo handshakes.â And then by the end of the first week of March, people were like, âThe world is legitimately ending.â
Alamin Uddin:
Exactly. Hopefully Josh is okay with me sharing this, it was Josh Buckley on that term sheet. COVID hits, and 100% of our customers at that point are dental offices, and you canât really go into peopleâs mouths.
Turner Novak:
Youâre literally just sitting there breathing on them.
Alamin Uddin:
And theyâre breathing on you. So they were all closed for the month. They werenât paying any vendors, werenât taking any demos, werenât signing up with any vendors. Nothing was happening. And on top of that, we couldnât collect from our existing customers. No new growth, nothing was happening. Josh is very high-integrity. We signed a term sheet, so he still followed through on it. But honestly, from my point of view, I was like, âWeâre probably going to die.â Because our market just disappeared overnight.
Turner Novak:
Yeah. So you also acquired a company that summer, right?
Alamin Uddin:
Yeah, hereâs kind of how we recovered from that period. Two things started to happen almost instantaneously. One is this concept of teledentistry started to come up at that time.
Turner Novak:
Teledentistry? How do you do that?
Alamin Uddin:
It kind of disappeared after COVID completely. Even I was kind of like, âWhy would you do that?â But if youâre a patient with an existing treatment going on, you still have to talk to your dentist one way or another every couple of weeks. So for those cases, either you come in or you do it virtually. New regulation passed. So what we decided to do was start giving away our product for free to anyone that would sign up, and we attached this video conferencing tool to it. And given it was free, versus our competitors, most of them were charging more.
Turner Novak:
They just increased their prices?
Alamin Uddin:
Yeah, because there was like, âWell, none of our customers are paying.â I donât know what their excuse was. But within a couple of weeks, every practice realized, âOh shit, we need to digitize really quickly here if weâre going to stay in business and continue to maintain our relationship with our existing patients.â So we went the extreme other direction, gave it away for free, and did that for like three to five months, with the terms that in six months weâll charge you again. Honestly, that shot up our growth significantly and gained us a lot of trust. We didnât collect cash from existing customers for like six months. And then, to your point, by the end of the year we did acquire a startup.
Turner Novak:
I think it was a pretty significant product expansion for you guys. Because you donât see many Series A companies make a significant acquisition as a percentage of the business.
Alamin Uddin:
Oh yeah, we spent half the Series A on it.
Turner Novak:
Really?
Alamin Uddin:
Yeah. In hindsight, itâs kind of crazy that we did it. We took a lot of strong, gutsy bets at that time, and nearly all of it worked out. I think we wouldnât be here otherwise. At that time our product was just one product, scheduling still and some patient communication. So if youâre a dental office five months into COVID, you can maybe start to see some patients for emergencies. Filling out paperwork by hand became a no-no as well, because itâs COVID, handling paper in general was a no-no. So customers needed digital paperwork, and there were a couple of vendors out there that did it, and a couple of our competitors already had it. Thatâs where they differentiated even before us.
Turner Novak:
Mm.
Alamin Uddin:
So that was a real gap we had. We could have taken the time to build it ourselves or partner with someone. The product was good enough, the team was good enough, a small team of like five or six people, so we just made a move and decided to acquire them and make it native to our product.
Turner Novak:
Interesting. And then you end up raising a Series B and a Series C over the next year or so. The company was growing pretty fast. Going back, itâs 2026, weâre halfway through 2026, was that a good idea looking back, to raise that much money?
Alamin Uddin:
Hereâs the fascinating part about NexHealth. Every time we raised capital, we didnât need it specifically, because remember, we ran out of cash at a certain point, so we had to run cash-flow positive by default. Then we raised some capital, but that lean efficiency, culturally speaking, never went away. So we were roughly break-even up until really 2023. But in between that, we raised a total of $176 million. So the question is why? At that time, there was a massive bubble going on as well.
Turner Novak:
Yeah.
Alamin Uddin:
However, at that time I didnât have a ton of VC connections to begin with. We were going from one fire to another every six months, first running out of money, then COVID happens. So I didnât really have time for it. So around mid-2020, what started to happen is we started to experience our very first couple of electronic health record systems trying to block our connection.
Alamin Uddin:
With updates. And that really panicked me. We ended 2020 with about 1,000 customers, and I was like, holy shit, if weâre only 1,000 customers and Henry Schein and some of these vendors own a massive part of the market, and they block us, weâre just going to die. So how do we solve it? What I realized is that the only way to solve this problem is to grow as fast as possible, meaning we have to have a lot of mutual customers with them.
Turner Novak:
Yeah.
Alamin Uddin:
And we have to do it so fast that these companies, which are really large and political, by the time they get their reaction together, weâve already snuck up on them.
Turner Novak:
Theyâve got a lot of meetings, and meetings about the meetings theyâve got to do.
Alamin Uddin:
Exactly. We kind of snuck up on them, so that was the goal. We went from 1,000 customers to 10,000 by 2023. The reason we raised was literally just to solve an existential problem, with an amazing product and value prop, so we knew it would grow. But for me personally, it was, âIf we donât grow, weâre going to die.â It wasnât even, âOh, youâve got to grow, you know youâre a startup.â It was literally, âIf you donât grow fast enough, there are five different companies trying to kill you at all times.â
Turner Novak:
Itâs not even that much capital compared to todayâs standard.
Alamin Uddin:
By todayâs standard, yes, especially today. Youâre right.
Turner Novak:
You couldâve raised that with, âHey, weâre AI healthcare,â and raised 100 million bucks.
Alamin Uddin:
Honestly, if I just changed our domain to .ai, I bet some random VC would throw half a billion dollars at us.
Turner Novak:
Their AI sourcing tool would service it to the investment committee and theyâll go.
Alamin Uddin:
AI is honestly so fascinating, because how do you even differentiate at this point if youâre a company, if youâre in any sort of B2B or even consumer?
Turner Novak:
Thatâs a good question. I feel like a lot of the way people answer that is to go into the physical world. That seems to be the solution.
Alamin Uddin:
Is the software category dead? Is it just Claude and OpenAI as the only software companies remaining?
Turner Novak:
But then at the same time, could I just say, âHey, Claude, go to nexthealth.com, copy it, come back in a day, make no mistakes, goâ? Could I do that?
Alamin Uddin:
You could do that, 100%. Well, it depends which part. But you could. Hereâs the really fascinating part about NexHealth. More recently I was having this conversation with an investor, theyâre not on our cap table, and weâre just joking around, like, âMan, if you were founded before 2022, youâre probably pretty screwed right now, right?â
Turner Novak:
Yeah.
Alamin Uddin:
Except honestly, the fact that NexHealth was founded before 2022 is one of the biggest strategic advantages we have today, just by happenstance.
Turner Novak:
Really?
Alamin Uddin:
Yeah, because we had enough time to go through all the pain of figuring out these integrations, to build the infrastructure where now, because of AI, within 18 months the API product grew so quickly, from 2024. The reason it grew so quickly is really because of AI. And the reason itâs because of AI is that if youâre building any sort of AI agent in our space today, the whole reason youâre building AI agents is because you want to complete tasks. Thatâs the value prop. You really canât complete tasks unless you can read and write data into those electronic health record systems, be it cloud-based or on-prem. And NexHealth is your only option today, because I donât think thereâs a direct competitor in our specific space, SMB, dental, dermatology, that has the number of integrations we have, the depth of data, the reliability. And the fact that you can go sign up and start building for free was a huge unlock for us. I have also never seen the rate of startups being created in our space.
Turner Novak:
Really?
Alamin Uddin:
It is phenomenal. Iâve never seen that before, ever, honestly.
Turner Novak:
So what are some of the most interesting startups doing in healthcare right now that youâre seeing?
Alamin Uddin:
The majority of the startups building with us, just to be very clear, one way or another have to do with AI these days. Of course, youâre building AI products. The really fascinating part is that on the SaaS side, we have direct competitors that are much bigger than us. They compete with us for scheduling, patient communication, and so on. Even they have very large deals with us to use our API to build AI agents and AI products for various workflows.
But to answer your question, whatâs happening in our space more and more is this. Thereâs a famous chart, I donât know if youâve seen it, administrative growth versus physician growth. Thereâs a massive amount of founders and engineers literally all hacking away at that problem.
Turner Novak:
And this problem is that the pace of jobs in administration is growing five times faster than the number of people delivering care?
Alamin Uddin:
Yeah, basically. I think the ratio is something like 50 to one, I could be wrong there, but itâs that big of a gap.
Turner Novak:
For every person administering care, thereâs an increasingly larger number of administrators around it.
Alamin Uddin:
Yep. And these administrators, for the example I gave earlier, itâs like two peopleâs job just to do the billing stuff. Someone has to pick up the phone, someone has to negotiate with the insurance carriers. A ton of this work goes into just supporting an office.
Turner Novak:
Mm.
Alamin Uddin:
So the more interesting ones, itâs not that the technologyâs fancy, but products are being built that you couldnât build before, because of NexHealthâs API but also because of AI, to overall solve the staffing gap these practices have. About 45% of practices in the country at any given time are missing one critical staff member that they need.
Turner Novak:
Really? Why donât they just hire someone?
Alamin Uddin:
Labor shortage.
Turner Novak:
Just pay more money. Do they not have more money to pay?
Alamin Uddin:
You also have to be trained. The latest data is that graduates have increased by like 20% over the last five years or so, but thatâs also not enough.
Turner Novak:
Probably attrition, people retiring.
Alamin Uddin:
Usually, for SMB practices, the profile thatâs working there is usually being paid hourly. They consider it a temporary, maybe two-year job, and then theyâll move on to something else, so the attrition rate is pretty high. You have to constantly replace them. So itâs a genuine problem for these practices and doctors, and at the same time, they canât function without them. Thatâs the gap being filled today throughout our platform, and I think thatâs really interesting.
Turner Novak:
So where do you think a lot of the value is going to accrue in AI healthcare specifically? Thatâs probably where you spent the most time. Where do you think the most valuable places to be building things are? Where do you think most of the value is going to accrue?
Alamin Uddin:
I can guess, and of course I have a thesis, and thereâs a strategy weâre chasing here at NexHealth as well.
Turner Novak:
What are you allowed to share about that?
Alamin Uddin:
Happy to share. Overall, in terms of sequencing, letâs consider the hype cycle generally.
Turner Novak:
So weâre in an AI hype cycle right now, youâre saying?
Alamin Uddin:
Exactly. And the different parts of the stack that are in the hype cycle. First it was the chip makers, Nvidia, in the very beginning. And then itâs OpenAI and Claude, the companies that do the R&D, the research to build the actual LLMs. And then I think the third wave, which hasnât truly come yet, but more and more of the ecosystem is realizing it, is that the third wave will be the companies and the infrastructure that takes the LLMs and enables them to actually action stuff in the real world.
When I say physical, what I truly mean is the layer that sits between the physical world and the software world, where these LLMs can then actually go into the physical world and action that data. A very basic example I gave earlier: if youâre a practice today, you can buy a product or build it yourself that uses the OpenAI API to build you a website chatbot product, kind of like Intercom.
Turner Novak:
In healthcare specifically, for a practitioner with the patients.
Alamin Uddin:
Exactly. You can go on the website as a patient and chat with it about your condition, and itâll recommend the right provider and slot in that office to book with. That chatbot or LLM needs to connect to the physical world, meaning what room is available, what day at that office in real time, because the patient is talking to that chatbot in real time. To then surface, âHey, hereâs whatâs available,â and take your information and write it back as well. Otherwise the patient will show up and it wonât be on their calendar. So if you just track the hype cycle: chip makers, the companies that build the LLMs, and then the infrastructure layer that sits between the LLMs and the real physical world. And of course, NexHealth very squarely sits in that stack. As a result, a couple of fascinating things are happening within our platform.
Turner Novak:
Yeah.
Alamin Uddin:
One is, nearly every AI startup. I recently went to this conference where I think nine out of 11 of the vendors showcasing were AI-native or built on our API.
Turner Novak:
Really?
Alamin Uddin:
Yeah. And we have direct SaaS competitors building with our API. And almost nearly every SaaS product out there is getting destroyed by AI-native products.
Turner Novak:
Really?
Alamin Uddin:
This is not about NexHealth specifically, I just mean in general. If youâre a workflow SaaS product, I donât know, a communication tool for lawyers or something, AI is really eating your lunch today if you donât evolve quickly enough.
Turner Novak:
Yeah.
Alamin Uddin:
The really fascinating part is that when it comes to NexHealthâs SaaS product, weâre actually not facing that pressure at all.
Turner Novak:
Mm.
Alamin Uddin:
And the reason is this. AI doesnât happen in our space today, and this is really not an exaggeration, I truly mean it, in the sense that we really do own the infrastructure layer for it. AI doesnât happen without NexHealth. So all of these companies, if they want to build AI agents, they have to build on our infrastructure. And they know we own these specific SaaS workflows, so they just donât go after it to begin with.
Turner Novak:
Oh, interesting.
Alamin Uddin:
Yeah, because we claim that. It would be like youâre building with OpenAI, but you want to build a chatbot.
Turner Novak:
Like a chatbot, like a ChatGPT type of thing.
Alamin Uddin:
Yeah, that would be kind of stupid. So thatâs the same effect weâre seeing here. All of this to say, thereâll be many, many valuable companies, and a lot of the developers building with us are massive, they have massive potential, some of them are growing way faster than we are. But where do you differentiate today? I think the most highly differentiated companies will be three types of companies. You can differentiate on distribution.
Turner Novak:
Mm-hmm.
Alamin Uddin:
You can differentiate on infrastructure, like payment processing or what NexHealthâs Synchronizer API is, or you can differentiate on some unique proprietary data. NexHealth has all three.
Turner Novak:
Hmm.
Alamin Uddin:
When I say distribution, hereâs what I mean. Every single time a developer builds any product on our API, they have to onboard some practice. If you go to Venmo and try to connect your bank account, Venmo has to onboard you to Plaid. So they have to do that as well, and that in itself gives us exposure for the SaaS product. Some practice will discover us from there, sign up for the SaaS product, and then be like, âWell, hey, what about my other vendors? Why arenât they using you guys? I wish they could read and write into my system as well.â Then some developer will sign up. So weâre seeing a really fascinating flywheel motion today.
Turner Novak:
Oh, where a practice will have one of their vendors use NexHealth, and then theyâll ask the other vendor, âHey, you should use NexHealth too.â
Alamin Uddin:
Exactly. Or theyâll directly buy our SaaS product and then realize NexHealth is the only vendor that can read and write data, and theyâll reach out to us like, âHey, how come my other vendors canât do this?â And a lot of the developers that build with us also want access to the practices on our platform. We own the infrastructure layer that can read and write data, and we have 89 million patients in our database. So we have really all three: data, infrastructure, and distribution at NexHealth today.
Turner Novak:
You probably had to be pretty patient though, because Iâm thinking ChatGPT launched in 2022. Were you just like, âLetâs do a bunch of AI stuff, should we train our own models?â What happened after ChatGPT launched? What was the evolution of how youâre thinking about the business?
Alamin Uddin:
One of the things about NexHealth is we generally like to wait a bit for things to mature before we start building with them. When ChatGPT first came out, within three months we had two AI features built into our SaaS product. And within six months, I flew our entire engineering team to Mexico for a three-day hackathon of just building a bunch of AI-native products and features. More recently, we also did another hackathon where you build AI products on our own API. So we were very early to it, but we actually never truly released anything user-facing.
Turner Novak:
Really? Why not?
Alamin Uddin:
Healthcare is one of those areas where being innovative and at the cutting edge is really important. At the same time, it has to be accurate, compliant, and it has to work. At that time, I personally didnât think the models were mature enough to support the types of use cases our customers would want to use us for.
Turner Novak:
So they would hallucinate some customer data and mess something up.
Alamin Uddin:
Yeah. Imagine, you have, what, three kids, I believe?
Turner Novak:
Only two.
Alamin Uddin:
Only two, okay, got it.
Turner Novak:
Some days I wish I had three. Some days Iâm like, âIâm just so glad I only have two.â
Alamin Uddin:
Imagine AI messes up one of their appointments or something.
Turner Novak:
Yeah.
Alamin Uddin:
And youâre a busy father or mother.
Turner Novak:
Yeah.
Alamin Uddin:
Youâd be pretty pissed. That is what often happens with a lot of the AI vendors in our space.
Turner Novak:
Really?
Alamin Uddin:
So itâs really important that the model is mature enough. And at the same time, what weâve waited for up until now is not only the models being mature enough, but releasing a version of our API, what we call an intent-based API. You can think about it as our own API that replicates what traditionally in AI-based products youâd call tools or workflows. Imagine we have 98 total endpoints. As a developer, you can weave them together and build workflows. Versus, we released a new version a couple of weeks ago where all of those workflows are pre-built for you, so whatever LLM product youâre building can just call those pre-built workflows and get going. So first, we want to take the infrastructure approach here, and then release practice-facing products. And you know what, weâre going to have practice-facing products coming soon as well.
Turner Novak:
Oh, really?
Alamin Uddin:
Yep.
Turner Novak:
How do you think about the fact that you have competitors building on your API? Thatâs a pretty sensitive thing to navigate. How do you generally approach that? Because someone might say, âWhy would I build on my competitorâs software?â People make the same argument with OpenAI, that theyâll just steal the best products and ideas. How do you navigate that?
Alamin Uddin:
Honestly, from my seat, itâs not like weâre forcing anybody. We have a really valuable product, and the market trusts us enough that theyâre doing it. So in terms of how you navigate that, I donât really think about navigating it, frankly. Theyâre a partner, theyâre a customer. Weâll give them a great service, build a great product for them, and enable them to innovate. Thatâs our value prop. And when it comes to competitive areas, sure, let the best product win. But for us, theyâre just as important a customer and partner as anybody else.
What NexHealth is, is a platform, and our mission statement is to accelerate innovation in healthcare. What we do not want to become is Henry Schein, where we treat different vendors specially or play favorites. None of it matters if weâre truly sticking to and being honest with our mission statement. The technology we have is amazing. It took us a while to build, it was really hard, but I think we have something magical here. And using it as a hammer or a tool to manipulate product quality or the market, the way the incumbents do today as the gatekeeper to innovation, would just go against our mission.
So from a cultural-principle standpoint, we tend to be very neutral there. And it shows up in the market, because these competitors, some of them very large public companies on the SaaS side, their CEOs trust us enough that they stake their entire business on our infrastructure. And generally, what ends up happening, like I gave that earlier example, is we do have our own lanes, everyone differentiates in some way. Because weâre the platform, it would just not make sense to go straight for the SaaS product and the core value prop we have.
Turner Novak:
Mm.
Alamin Uddin:
There are plenty of ways to differentiate and go after workflows, and the market is large enough. So more organically, competitors end up staying away from our core lane over time.
Turner Novak:
So itâs really like the whole OpenAI and ChatGPT thing from earlier.
Alamin Uddin:
Exactly.
Turner Novak:
And you donât really think of yourself as a healthcare company. You guys specifically say, âWeâre a technology company.â Why is that so important to how you think about things?
Alamin Uddin:
Earlier we talked about why it worked with NexHealth versus every other startup that tried before this. One of the core things I realized, maybe around 2022, is what everyone else did before us that was different. It was all relationship-based. Like, âHey, weâre going to go sign partnerships, have good relationships with you, and thatâs how weâre going to solve this problem.â Versus we took a pure technology approach. We said, âHey, this is a technology problem. Weâre going to build some unique technology that can sync data back and forth, make it reliable, stable, and scalable enough that it can build massively scalable products on top of this infrastructure,â from Invisalign on down.
So healthcare, and more generally nearly every space, taking the approach of building great technology and great products is the way to change healthcare, versus anything else. And itâs really important now for us to differentiate that way, because the majority of companies, even if theyâre tech companies in our space, culturally, in the way they describe themselves, will say, âHey, weâre a healthcare company.â And that has a ton of consequences, because the way you approach problems is different.
Turner Novak:
Mm.
Alamin Uddin:
Again, the example of, do you partner with these EHR systems, do you pay them 30% or not? Or do you just treat it as a technology problem and solve it with technology? Depending on how you think about yourself, you make a different decision. It may not feel that important from the outside, but at least in the world of healthcare, and especially internally, I really emphasize it every couple of months in all-hands: âHey, weâre a technology company that just happens to be in healthcare solving problems.â And I think if we continue that approach, weâll continue to succeed as a company.
Turner Novak:
And you donât really hire from the healthcare industry. Most of the people in NexHealth, their background is in technology. They worked at Netflix or Databricks. So that feels pretty important to keeping that ethos.
Alamin Uddin:
Yes.
Turner Novak:
How do you convince people to come join? Whatâs generally the pitch for joining NexHealth versus, you know, people working at Anthropic right now, though I guess only a couple thousand people can. Whatâs the pitch to convince people to come work there?
Alamin Uddin:
Itâs twofold. One is we have a great mission. In addition to that, NexHealth is really for a very specific type of person and a specific type of culture, where not everyone is the right fit, to be very clear.
Turner Novak:
So whoâs a bad fit? Or whoâs a good fit?
Alamin Uddin:
Just to be blunt, a bad fit would be someone that would join OpenAI today, or Anthropic today.
Turner Novak:
Really?
Alamin Uddin:
Yeah. NexHealth is for the type of person that genuinely wants to have real impact on the world, that genuinely wants to do good for the world more directly. But theyâre also so ambitious that they want to build something that will outlast OpenAI or any other software vendor I can think of. Hereâs what I mean. Healthcare, and the lack of innovation and this problem statement, is a type of problem that every single human before us has had, and every single human after us will have. So itâs really a species-level problem. And the ambition we have here, building the infrastructure where innovation happens, if we can unlock that, I think weâve changed the course of humanity, versus anything else, frankly. And healthcareâs so personal. Moneyâs important, work is important, productivity tools are important, all of it is really important.
Turner Novak:
Mm.
Alamin Uddin:
But the one area where it matters the most is everyoneâs health. If you can impact that, and at the same time scale it globally, build something valuable, institutionalize it, make a company thatâs generational, thatâs our ambition here. If you can do that, then what I like to call, youâll be a legend.
Turner Novak:
Whatâs the importance of shoes at NexHealth?
Alamin Uddin:
Oof. Itâs a symbol of being scrappy. This comes from, youâll recall I hired this college kid, Bobby, to be an SDR. It was his first job.
Turner Novak:
This is the first one.
Alamin Uddin:
His first job ever, one of our very first hires at NexHealth on the sales side. At that time I was closing all of our deals, and I was like, âAll right, maybe let me hire someone that can create pipeline for me.â And of course, he was ambitious, and within three or four months he wanted to start closing his own deals. So at that time I was like, âYou know what, how about this? As soon as you close your first deal, Iâll get you a gift.â And he did, and I walked over to the Allbirds store, spent like $60 on a pair of Allbirds, brought it back and gave it to him as a gift. Since then, thatâs a tradition that has continued at NexHealth, where when you achieve a milestone, that first milestone, everyone just gets shoes.
Turner Novak:
Was he literally walking around closing the deals when he was the BDR back in New York?
Alamin Uddin:
Let me think. By that point, it was like 50/50.
Turner Novak:
Okay.
Alamin Uddin:
Because I was like, âAll right, maybe you donât have to go into the office.â
Alamin Uddin:
But anyways, yeah, it was 50/50. At that time we were doing anything and everything. I would drive two hours to go to an office sometimes.
Turner Novak:
Yeah, you were in Detroit for one of those big customers. And you guys, I think you donât use Slack as much as the average startup. You do a lot more email. More structured in that way?
Alamin Uddin:
Thatâs my desire, as the CEO, but thatâs not true, honestly. Slack is just messy, I donât like it. Company coordination, internal communication, from the tools to how you effectively coordinate a set of people cross-functionally and then execute. And when youâre doing the execution right, itâs generally something new. Maybe itâs been tried in other spaces but hasnât been tried in our space. Great example is the payments product. Youâd think Square knows who paid this bill, but those basics donât exist. So we do have to invent new concepts in our space.
Turner Novak:
Yep.
Alamin Uddin:
So one of the reasons you see me complain about internal communication tools in general is that coordination of speed internally is just so important, especially when a lot of the people we have come from outside of tech. And on top of that, itâs a problem weâre addressing that no one else has usually solved yet, and itâs a new thing for the market. A lot of things just have to go right, honestly. Execution is really hard but really important.
Turner Novak:
Are there things inside the company youâve changed because of AI?
Alamin Uddin:
The obvious one is, something like 100% of our code is written by Claude these days. Hereâs a change, at least for me in my thinking, and itâs taken a bit of adjustment in two areas. One is because of how advanced Fable is. Fable 5 is just freaking amazing. Because of how mature these models have gotten, you can genuinely say, âHey, Fable 5 is probably close to a staff-level engineer today, or will be very soon.â So what does it mean if you take that same skill set to marketing, legal, and so on?
I donât think it necessarily means you hire fewer people. I donât think thatâs the right way to think about it. But the way you get work done fundamentally changes. And if you donât change the way your teams are working internally, youâll be left behind, and someone else will out-compete you really easily. Most people look at it from the lens of, âHey, itâs about the number of people,â as the ultimate end outcome. Honestly, I donât think thatâs whatâs going to end up happening. I think youâll generally have the same number of people, but the same number of people are now getting things done at 5x the speed and maybe 10x the output. No company in the world will say no to that. Thatâs whatâs happening at NexHealth as well. To the point that, I am not even kidding, a random problem weâre having internally the last couple of weeks: our marketing calendar for product releases doesnât work anymore, because there are just so many things to ship.
Turner Novak:
Oh. Okay.
Alamin Uddin:
Yeah.
Turner Novak:
When you look at the line-item cost expenses in the budget, what are the biggest AI tools you guys are spending on?
Alamin Uddin:
Weâre on both OpenAI and Claude. Weâre sort of model-agnostic, but for different use cases. For product-facing, we use OpenAI. For internal, Claude, but Gemini as well. Our VP of engineering, Rahmi, spent twenty years at Google early on, then Databricks, and heâs also a PhD in machine learning. So he has a lot of opinions and a lot of expertise, and understands all the pitfalls. So we tend to use all three, but it all depends on the team, and I think itâs important that any company remains model-agnostic.
Turner Novak:
One last question. Do you have a favorite founder, CEO, or business from history that youâve gotten a lot of inspiration from?
Alamin Uddin:
This is going to be surprising for a lot of people, but for me, frankly, itâs Microsoft. The reason is that Microsoft was one of the very first companies out there to scale and really invent the business of software. Before Microsoft, software culturally was expected to be free. You paid and charged for hardware, but the software was a hobbyist thing. It was free by default. Itâs not that valuable, basically.
Turner Novak:
It all kind of sucked, compared to today.
Alamin Uddin:
Thereâs not much you could do with computer software back then, technically speaking.
Turner Novak:
Fair.
Alamin Uddin:
But anyways, Bill Gates was the very first one to start monetizing software, and everyone kind of hated him from that point onwards.
Turner Novak:
Really?
Alamin Uddin:
But the reason I admire the Microsoft story specifically is that at NexHealth, it truly does feel like weâre doing something very similar. When I say weâre inventing a lot of new concepts and new products, new ways of doing things, and also fighting a ton of inertia, and creating this ecosystem for innovation on top. The best company I can think of from a business model perspective, to the number of products, to the strategy, and really everything, is Microsoft in the 1990s. Iâm also a huge history nerd. Microsoft â77 to 1990 would probably be my biggest inspiration.
Turner Novak:
Do you have any favorite people from history? This could even be 10,000 years ago. Caesar, Cleopatra, George Washington, anyone interesting. The thing Iâm always fascinated by is, youâll read this story of the American Revolution, like we just randomly were like, âHey, letâs break away from the British,â and we won a war. But it had to do with a business, commercial reason for it. We didnât just wake up and say, âHey, letâs secede.â There were incentives, and there was a whole business behind it.
Alamin Uddin:
Mm-hmm.
Turner Novak:
Thereâs a pretty interesting podcast I listen to. My friend Ty Morris has a podcast called Relentless, and he talked to Ryan Petersen, the founder of Flexport. It was 45 minutes of just talking about the history of logistics. And it was fascinating that every single big historical event you think of was related to some kind of logistical problem. How global trade evolved, and the economy evolved around how a new technology enabled a boat to go farther, or a wheel enabled the trade route to be longer, or planes enabled trade in different ways. Iâm always fascinated with the history of how these non-obvious things really impact stuff.
Alamin Uddin:
All of history is driven by trade. Really, all of it is driven by trade. Trade routes, ability to trade, and money movement in a sense, where does the money go, goods. Hereâs the thing, though. I am a history nerd, and over time Iâve evolved my point of view on a lot of things.
Turner Novak:
Really?
Alamin Uddin:
One is the admiration of great men, in a sense. Iâm not one of those people that donât believe historyâs driven by great historical figures. I think thereâs much more nuance than that. But at the same time, what I admire more are the structures, the institutions, as well as the conditions that exist to drive some change, and the people that saw that opportunity, took advantage of it, and then drove that change.
Turner Novak:
Wow, this has been a lot of fun. Thanks for coming on the show.
Alamin Uddin:
Yeah, of course. Thank you for having me as well. Honestly, a really fun conversation here.
Find transcripts of all other episodes here.
