Sridhar Uyyala - CEO @ TensorLinks

Welcome to another Founder Wisdom podcast. We have, uh, Sridhar Uyala, which was on my Spark Tank show a couple of, uh, days or a week ago. He's the founder at, uh, CancerLynx. Uh, we're gonna talk about, uh, his business, an AI front desk, uh, agent for dental clinics. So Sridhar, welcome to the pod. Tell us a bit more about yourself and what you're up to.

Hey, Charles, thank you for having me here. And I had a great time with Spark Tank last time and really appreciate it. I'm really excited to be here and talk about my business and answer any questions. Thank you.

Yeah, tell us a bit more about your background and yeah, what led you to TensorLink?

Yeah, I'm an engineer, mainly a software engineer, a bit of a research background. And I did a lot of technical work in, in my career, uh, mainly on around the operating systems and the file systems, uh, and deep learning, uh, computer vision, things like that. Um, and then I started a company with the idea that the AI will be here, 2017, to build an assistant for the cars. Again, over the years we kind of pivoted, uh, and came back to healthcare where I thought, um, I mean, we're kind of now making a ways where Making it easier to run your clinic, reducing the administrative burdens. So if you're a clinic owner, if you have multiple clinics or single clinic, doesn't matter, we know that the administrative burdens are more and it's a manual process running them, requires a lot of staff and also inefficient. And we are bringing our AI that's gonna be embedded with your staff and take on some of those administrative work. And it, it's as simple as that.

Interesting. What's the traction so far? What's, uh, how many clients you have and how much you're raising?

Yeah, we have, uh, 27 clinics we are live at, uh, today, as of today. And we're trying to raise, uh, our pre-seed, uh, with $850K. And we are bootstrapped.

What's your GTM system? How do you get to all of these clinics?

So far it has just been totally Google Ads and ChatGPT. Some of it is. And right now we are just doubling down on those two channels, Google Ads and the ChatGPT. And also we now just started to look into LinkedIn and social media as one more GTM and cold outreach over the emails with the customers that we want to do. Mainly showcase them what we can bring, how we can transform their clinic administrative operations. Just the conversations around it and being visible that we exist.

Your competitors, right? There's a lot of them in that space specifically. How do you— what's your little moat when it comes to them? Like, how do you outpace them?

That's a good question because a lot of the people ask the same question. Like, hey, there's like a lot of, um, the competition in this space if you just talk about a receptionist. But we don't position ourselves as that way, just a way to like initially communicate to our customers is that, hey, this is a receptionist that's gonna automate some of your workflows. And when you come to competition, you can see that you can Google search and find most of them say it's a voice assistant or like a voice bot or a chatbot that's gonna help you book appointments. I mean, that's not how we see it at all. Like we are completely like building an operational manager per se for you to run some of those operations. Right now we just started our patient engagement like around taking calls and not even just taking the calls. I mean, we actually, We treat voice as one of the communication channels. That's it, period. And we actually support, you know, right, if you're hiring a staff or an operational manager, what do they do? They're not just gonna communicate on the phone, right? They would run all of the operations. They might be checking your emails if the patient are asking and they would check the SMS and they will respond, right, as staff. And also they would answer any set of the questions and they will manage the schedule for the doctors, providers. And they would look at treatment plan and say which patients to reach out if they're not showing up. There's a ton and also insurance verification application, submitting claims, billing. There is a ton of administrative work, and we don't see that it is just a voice bot. We're just actually thinking an operational layer, and it just so happened that we just started on, on the receptionist workflows.

Interesting. And yeah, tell me about the dentist. When you meet them, what kind of conversation do you have with them to qualify them for your products? What problems do they face? What are the, the fires burning in their offices?

I think it mainly goes back to, some of the conversation goes back to how do they manage the calls? The 30 to 40% of the calls go missed just to look at one of the problem, right? And then staff need to be available during the after hours and it has to be consistent, right? And book the right appointments into the right rooms, right? Other than availability, and also not only that, like some of these clinic owners that we had conversations with, they service different demographics. Like you need to be speaking multiple languages. Which is really hard. Again, you have to find the translation services, speak to the— otherwise you're losing the care for those patients. So that's what the Bridging the Gap is as an initial footstep. Like, hey, our AI can handle if you guys are busy, all the calls, and book directly to the appointment the way you wanted it. It's not like we capture the details and send you an email or something like— not like that. We are fully integrated into your systems, right? Our AI understands the way your operation manager is understanding, uh, your practice, like including what providers do you have and which rooms they are assigned to, uh, and what, what kind of insurances would accept and where would you, uh, and what kind of appointments that you book them for. If you are— if I'm a new patient and coming into your practice, like how many— what would you do with it? Would you do like a full oral exam? Or if, if I have a toothache, I need to see someone emergency, What kind of an appointment would you do? Would you just do a full scan or would you just do a limited exam? Things like that. Our AI is embedded with that operational knowledge so that it is acting like that. We educate them saying, hey, this is not just a bot that's going to collect and add more work for you. Instead, actually it's going to reduce the work for you by understanding. The main thing about ours is we don't have to change the way you are running your operation so far. This is going to be like a helping hand that is going to seamlessly get embedded, and that's it. You have the full control on when the AI should barge in and what scheduling parts it should take care, and you have full customizability with our platform.

Interesting. Do you focus on some specific territories first when it didn't make sense to deploy in Fresco first and in Frisco, that is, and then in Austin, and then in Houston, and then like full US?

Not really, to be honest. Of course, we look into our concentration of the clients where they are, major metropolitan states and cities, right? You know, the New York has a major concentration of the dental clinics, New York area, New York state, and then California and Texas, Florida, where there are a lot of dental clinics. But our software is like cloud-based and it can service anyone in North America. In terms of the GTM strategy-wise, obviously I live in Dallas area. I would want to actually, if possible, physically reach out to any of the decision makers here and set up meetings with them. That's like one advantage I have. But, and also some of the, the other states like California, where the technology is a little bit more adaptive for the AI. So we could see a lot more traction there too, comparatively, from other states. So there is no— and that's why we are spending more on the digital marketing than the regular sales, like, hey, there is region and stuff. And also the AI is in the very beginning phases, meaning they know that the AI exists, but not many are ready, and they don't know what it means to bring AI and deployment to their workflows and into their clinic. They're depending on the traditional system so far and there is always a hesitance and inertia to get a new technology and they worry about the security, they worry about the compliance and the vendors. They would have to work with one more vendor now without knowing what they bring to the table and the IT would raise the issues like, hey, we don't know what they're bringing in. There are a ton of things that would go into deciding which customer would try to adopt first and that's why instead of a very specific region, we broadly focus on North American, where metropolitan cities. And organically like talk to someone like who already kind of coming into us. But that conversation is gonna slowly change and so does our approach too.

Got it. And the 25 clinics, they pay you monthly? Like what's the deal with them?

Yeah, yeah. They pay me, 27 clinics right now, but they pay me monthly. It's a monthly subscription-based, very simple model. Actually, we publish our pricing also on our website. And we have like a tier like, hey, if you just want a receptionist that is gonna, you just want a voice, nothing else, call management is your problem, we have a solution for it. Or you just wanted to do a full engagement suit like across any channels and we have just that tier for it. Or you're just looking for more, even like looking at your schedule and proactively reaching out to the patients outbound as well, that we have another tier. So we just like simple 3-tier model and the price according to the market and it is a monthly subscription. That's what ours is.

And you said ads, have you been cold emailing or cold calling these clinics as well? Have you tried outbound?

We were just trying the cold email as we speak, like last 2 weeks or so we just started. We used to like email a little bit, very focused, not like a full GTM kind of focus. Now we kind of realized understanding like who our customer is, like where would it benefit the most and trying to reach out to those customers over the email. Mainly not just as like a sales push, but more like, hey, we see that mainly showcasing the results. Like, hey, we are with these clients and these are the value add that we're adding and this is the traction AI is enabling. And I would love if they would consider.

Interesting. This year you want to raise, right? How's the fundraising going and why is TensorLinks more attractive than other types of investments that would be pitched to VCs?

If you're just looking for another chatbot, I would say don't invest in us. But if you really think that we are building a serious platform that's going to make an impact into healthcare, that who would understand the operations end to end and will show us the results and traction, I think you would definitely want to invest in us. And that's why we're very attractive. And not only that, we are actually, we're able to manage this bootstrapped without raising any money so far. And we internally automate our stuff, like using a lot of the automation tools and AI. So all the processes, we treat ourselves as like an AI company. That means that we can bring a lot of efficiency. So if you're investing a dollar into us, we are building a future company, a native company for dentistry, and it's going to make a lot of impact in other verticals as well, especially in healthcare. So yeah, if you're an investor watching this, I would say we are worth every penny and we have a lot of traction and we are very scalable.

And give me an overview of the team.

Yeah, so I would say our team is put together very organically, like people who are quite a bit interested and want to get their hands on AI and quite excited about the problem they're solving. We are all engineers, I'll tell you that, except the one person who would help with the operations, even understanding the clinical operation. Some of the administrative side of things. Other than that, we're pretty much all AI engineers. I, being an AI engineer, would have a lot of— Can you hear me fine? Yeah, we can hear you. I'm sorry. Just to give a bit of a composition for you, we have engineers that focus on our agent framework and we have engineers that would just focus on the compliance and the security part of it, like HIPAA and SOC things like that. And we have engineers that would just focus on the infrastructure and then engineers that would focus on automating the rest of the flow, apps, things like that, and the cloud-based infrastructure, things like that. And that's our composition. And now we are slowly wanting to— now that the product is there, we kind of built a very good product where we're confident. Now just want to build other things like the GTM engine and marketing and sales slowly.

What's been the most challenging part so far?

I would say many things. I would start with the product side first, just to touch the base on, you know, that the customers would come with a lot of expectations when it comes to AI because they hear in the media that the AI is going to do wonders, right? But again, some of the technologies, like it is, it is actually, it is a transformative technology, But some of the channels like the voice and stuff, in the very early phases, it is still rough. There is a lot of challenges to get the voice right. Let's say you have a patient that's going to speak a little bit slow or a patient that's going a little bit faster, or you're spelling out your names, it may get wrong, or you're giving your date of birth or a bunch of numbers that would include insurance ID and things. Getting those right is crucial. Some of these models are getting better at these, like from transcription to capturing. But actually there is a ton of engineering work that goes into to get these right. It's not like as you see in the demos that, hey, receptionist takes the calls and everything goes smooth. No, it can hallucinate sometimes and pass wrong date of birth, which is getting better over time. But we have gone into a lot of issues. Perfecting this engineering, even though there is a lot of AI in it, to get these models work for us in the production scenarios. We even created feedback loops that we look at how things have happened every day. Again, this is automated and see how we can change to make it better and guide the models better. Not only that, the guardrails, what can be done to follow the instructions correctly and not get into other things that are not intended. Adding the guardrails, making security, and follow the instructions is really crucial. And even like, I'll tell you a simple scenario. If someone elderly is speaking to an agent, they might be like really slow in saying things. And sometimes our agent has to like figure out, is the person still there? Do I need to say, are you still there? Right? Sometimes it might barge in very fast and sometimes people are really is speaking fast, it has to keep up. So there are like a lot of engineering challenges that were not seen, uh, but we kind of work on those like tough problems on the product side. And as you know, like, we're engineering background, we have built this, but we don't— we didn't figure out the right GTM strategy or anything yet. That is another challenges that we have, like how do we make sure that even though we started very early compared to everyone else in the market here, uh, 2023 end we started on this project on the agency platform. I built it from the scratch as an in-native way. Now there are a ton of companies out there, but being ourselves visible and able to differentiate is one of the challenges we feel. How do we communicate that message? For example, this podcast would really help set the ground. Someone could just get the website and only look at a few minutes. Oh, this is another AI risk company. But if someone is listening to this, they would understand what this means. What are the hard challenges? And what it means, what can we bring to the table. That's why I, I'm really excited to speak to you on this.

Well, yeah, it's— I think it's a, it's an intriguing company. Um, you said you, or your LinkedIn says at least like 10 years, right, that you've been doing this. I think you accepted the job in between. Like, are you full-time into this?

Like, what's, uh, why did you accept that job and not go all in back Yeah, as I mentioned that I started this company in 2017 itself with the goal to build an assistant for these cars, hoping that the self-driving cars will be here, things like that. 2017, the voice models were not there and the interactions wasn't good. And it's the first time for me to figure out with an idea to how do you take it to full market. I was just fooling myself, like, thinking that how big of a problem versus like how easy it is tractable and if the tech is there or not. I quickly realized it within the 8 months that, that I didn't understand the fully the gravity of the problem and how to solve it. So then I, again, I had to like go back to job to support my living. And I happened to just land onto CVS Health, which was like a very good, interesting opportunity into healthcare where I worked on the patient engagement piece, like digitizing the patient experience for the CVS Health customers. And that kind of gave me, I was always like soul searching to see like what I would do, like when the AI would come or like when would the AI be. And as I noticed, GPT-2, GPT-3 came, model in 2022. I, I took the bet again, left the company, left the, the, my job, and did work full-time from there onwards. Uh, it took a year to figure out, uh, come to where I am under the dental space, uh, where the, the stack that I was building gonna make an impact. Uh, but it— that's why the journey is so high. But dental, we started on 2023, and that's why you see like a 10-year-old company. But really, I was like going through a lot of, uh, pivots to like figure out like what would you do and where would I apply what is the problem. And I'm new being to healthcare, there's a lot of challenges for me to like come and understand the, the domain and the decision makers here and the leaders, how the industry work, right? It is, it is quite a bit of challenging and upkeep and a lot of learning, and I'm really happy that I took this path.

Interesting. Well, Siddharth, thanks for showing up today. Where can people find out more about you and your company?

Sure, I mean, you can reach out to me at TensorLynx— sorry, sridhar@tensorlynx.com— or go to our website, uh, tensor tensilinks.com. And you can reach out to us on even on X.com with the Tensilinks.

It was Sridhar and Charles Cormier.