Six New York AI Startups You Need to Know

Transcript generated automatically from the episode audio and lightly cleaned up (filler words removed, misheard names corrected where clear). Speaker labels were added afterward and may contain errors.

Rob May: All right. Hey everybody. I'm Rob May, the co-founder and CEO at Neurometric.

Ryan Eppley: And I'm Ryan Eppley, co-founder CEO of RootAccess.

Rob May: And this is the AI in NYC show. So today's show for the end of the year, we shot this a little bit early. We're going to post it the last week of the year. And what we're going to do is we're going to profile some of the interesting AI companies in New York. A lot of people don't know this is sort of the applied AI capital of the world. Hundreds maybe thousands of AI companies here. Definitely hundreds that I know of. So we have six or seven of them today and we're going to ask them some questions and ask what you can do as an AI in NYC viewer to help grow these companies so that we can continue to be the applied AI hub. So our first company today is called CounselPro. My former colleague actually Ian O'Brien's the CEO. So Ian, welcome to the welcome to the podcast.

Ian O’Brien: Thank you. Thank you very much for having me, Rob.

Rob May: Yeah. So tell us a little bit. So give us the overview quickly like what does CounselPro do you know how old is the company?

Ian O’Brien: Sure. So the company is a year and a half old. We've started out with this hypothesis that we could take unstructured data and make structured data out of it which really isn't a new idea. But what's different about it is how we apply that. And so we apply it specifically to the legal practice. Now the reason why this idea came about actually because I went through a divorce during COVID. And so we were separated for a couple of years and you know things looked like they were going to be fine, but then as soon as it came time to have a divorce, suddenly, you know, one trying to get a whole lot of money. And so, what I wound up doing is I couldn't afford to have a forensic accountant come in.

And so, as part of a divorce, you have this thing called mandatory disclosure. They have to give up all their bank statements for like the last three years. And so, I went line by line. I copied every date, every description, every amount. I Googled it. I was like, what, you know, what was this spending? I put I categorized it. I made some charts like I would for a marketing meeting and I handed it to my attorney for deposition where he used he basically just went through my data when when he deposed my ex and wound up reducing the settlement demand by 85%. Wow. And he comes to me and he says that was incredible. Like all I did was use your data. Would you want to come and do this for my firm full-time? I said no absolutely not. That sounds miserable. But I figured, you know, light bulb went off. This is a really boring problem that a lot of people have that I think that nobody's really solving today. So maybe I can come up with a solution that solved it and it just happened to be at the same time that AI was taking off. And so the available tools to be able to build it quickly, have the infrastructure running without having a whole lot of capital expense was was huge for us.

Ryan Eppley: So does everyone on your company have to be divorced or go through it? Is that a like prerequisite?

Ian O’Brien: No. No, they don't have to be divorced. And in fact, more than just divorce attorneys use it. That was the original hypothesis. But essentially what we do is we take bank statements, credit card statements, any financial document, scan all of it, pull it into a database, categorize it, and then we have an AI forensic accountant that runs through it. We do a lot of really cool stuff like we do like fancy math with with IQR range so we can figure out whether or not spending was above an average for a specific category. We can start there's a lot of times people ask where did all the money go? And so we can within only a couple of minutes have a flow of funds diagram that shows exactly where where all the money was transferred to names all those parties. So you can imagine you know bankruptcy really loves this as a use case. Consultancy really loved this as a use case. So it's really expanded far beyond who we thought our original audience was going to be. And the awesome part about it is it's completely organic.

Ryan Eppley: Yeah. Yeah. Fantastic. Congratulations. What part of the city you in? Because you're here in New York. Where where's the offices?

Ian O’Brien: It's all remote. We have no offices. It's part of the ways that we can, you know, we can cut costs and continue to be bootstrapped is we don't have an office that we have to maintain.

Ryan Eppley: So, where do you live then? Because that's the office.

Ian O’Brien: Well, where I live is actually in Westchester. I'm a cheater.

Ryan Eppley: Oh, great. Close. Yeah. Yeah. We'll call we'll call it New York.

Rob May: All right. Great. Cool. And so tell us a little bit about, you know, the stage of the company. So, you mentioned your bootstrap. Yeah. Are you guys generating revenue yet? Do you have some customers? Where do those customers mostly come from?

Ian O’Brien: Yeah, sure. So we have over 50 customers today which comprise small and medium law firms. We also have several cities that use us and we also have some larger consultancy companies that use us for data ingestion and extraction. So we're we're far beyond pre-revenue. We're we're we're generating great revenue today. And it's all organic completely organic growth. And it's mainly because we can do something that no other system out there can do today. It won't always be the case, but we are early to market for a product like this. And also I have an SEO background in marketing. And so I write a pretty decent website, pretty good marketing copy, which thank you Claude for the help in writing my blog post. Yeah. Has resulted in getting picked up by all of the LLM. So when somebody asks ChatGPT, you know, how do I do this? CounselPro comes up.

Rob May: That's awesome. There's whole companies now specializing in that, helping you get found by the LLM. So it's really good that you guys have hacked that in the beginning. And tell us a little bit about the AI structure that you use is you mostly use calls to LLMs or using any proprietary models that you build like what's the technology like in the background?

Ian O’Brien: Yeah, so one of the models that we use is a modified open source model that's actually built here in New York City. It's by data lab. It's called Chandra. It's wonderful. So we use part of that to extract. There's a little bit more than just doing an LLM call to that. So we have to do inference. We have to figure out how to scale that across a wide customer base that's very bursty with demand. So it's a little bit more complicated than just calling the LLM. And then we use some open source models to categorize that. Right now what we do use the APIs for is we use Gemini for our forensic accountant mainly because well it's just so good. The context window is huge and the comparisons it's able to pull up with just these these very random events that happen within transaction data. It's able to correlate them in such a almost truly magical way that seeing that applied where it doesn't have to write contracts, it doesn't have to do math, it just has to draw similarities between two things which AI is exceptional, really good at.

Ryan Eppley: Yeah. Yeah. You are you hiring for any more roles? Growing the team at all for our viewers to maybe think about.

Ian O’Brien: We just hired for the latest role. So, we currently have three engineers, one forensic accountant, and then there's me. So, no open roles today, but we're hoping that that'll change sometime pretty soon. We're trying to keep things as lean as possible so that we can stay as dynamic as possible as a team.

Ryan Eppley: Cool. And then for viewers, your camera's right over there. Where where can they find you?

Ian O’Brien: It's CounselPro.ai.

Rob May: Fantastic. Well, thanks for coming on the show, man. Appreciate it.

Ian O’Brien: Oh, thank you very much.

Rob May: All right, so next up we have Alec Glassman, the CEO of Silvershield. Alex, welcome to the podcast.

Alec Glassman: Thanks. Nice to be here. Thanks for having me.

Rob May: So, AI is a new technology widely adopted by young people. You are probably our only startup that is going to focus on the elderly. So tell us a little bit about Silvershield, what it does.

Alec Glassman: Yeah. Well, first thing by the way, before we even get into that, I think it's super interesting. AP has some great research on this, but adoption specifically with AI with older adults is really really high.

Rob May: Oh, that's awesome.

Alec Glassman: And I think in general there's and I've been like it's been cool to learn this as I've been getting deep in the space. But like I think we underestimate how older adults use technology. And it's been really interesting to see that like they are us but a little bit older, right? And so like sure maybe we're technative but actually they they really use it a lot and so that's been really interesting and some fun insights. But yeah, I'm Alec, founder and CEO of Silvershield. We make tools to make it easy for anyone to spot and stop scams and fraud. And we're focused on helping older adults to start because they're impacted the most by that.

Ryan Eppley: I actually got to see you pitch I forget how many months ago it was but you know you said how you did some field testing right like walk us through like you actually being in the field like would you call yourself a for deployed engineer in that sense like in these home care all Alex's best friends are now in their 80s right right walk us through how how you kind of got this up and going

Alec Glassman: well so it's been interesting and it actually progressed since then a lot and I had some more cool experiences there and so yeah so basically when I first started with Silvershield. The first thing I did was I'd live with my folks in a retirement community in Florida for 2 months and I spoke to older adults every single day. And it was interesting because I think that first of all like I had an idea in my head. I know Rob, we had talked about this a lot about what we thought the problem was and what we thought the solution was. And you know, talk to your customers, you talk to people, you learn the issues. And I realized like there's a different approach here that I think is more impactful and makes more sense. And the other things I learned was just how big of a problem scams and fraud are, how misunderstood it is. The whole human side to it that I think people really forget that's really important.

And so with that experience, basically when I left after those two months, I had built a new prototype. I built it myself on like bubble on my phone and I had gone around and showed a lot of people and then I basically spent the next few months going around the country to different conferences talking to hundreds of you know professionals you know ex FBI ex CIA exNSA adult protective services case managers older adults themselves of course and like people were really really excited and I would show them this thing and they'd be like that's really cool can you do this that's really cool what about that and they kind of kept enhancing and enhancing and next thing you know I had a commercial pipeline and I'm Okay, we got to get a CTO and build a real version of this. And so I did that. And then in August to kind of cap off that experience, I was invited to live in a senior living community in rural Ohio. And I lived there for two weeks and that was just an awesome experience.

Ryan Eppley: The only person that was allowed in the over 55 community. You know, congratulations.

Alec Glassman: Yeah, it was cool.

Rob May: Yeah. So, when did you start this again and where are you today in terms of your attraction?

Alec Glassman: So started in February and then in terms of traction so some really cool things that are happening right now. So in January we're launching across New York State with the state and we're really really excited about that. We have a couple of partners in Pennsylvania. We have some folks that are running some testing for us out in California starting in the new year and partners in Massachusetts and some other really cool things that I probably can't talk about yet coming down the pike. But, it it's been cool and like part of what's been really cool is, and this has been a recent thing, and I guess y'all are used to this because of what you do. But we've had the support of the public and the media behind us recently, and that's been really interesting. So people have picked up on this. We don't do any PR or anything like that. We're focused on like building good solutions and helping people.

Ryan Eppley: Yeah, helping grandma not get scammed. Like, who's not going to support that, right?

Alec Glassman: It's important and I have to think another like little interlude here that's really important. Scams and fraud are an everyone problem and there's been a lot of really good research that's comes out that's come out recently where it turns out younger people are more likely to report losing money to a scam than than older adults are. But the results for older adults are just more devastating. If you lose your retirement savings when you're out of the workforce, that's a huge problem. And so, and so that that that scale is huge. But it's really interesting like since I've been in this space, people find out what I do and then they've shared things with me that they don't share with other people because they don't want to talk about it. People of all ages. So I think a lot of people have this problem and I think that people are excited to see someone really trying to tackle it and recognizing that this is a human problem. There's a there's a person being affected by this.

Rob May: And at a high level, how does the app work? Is it voice based? Is it text based? Is it on your laptop, your mobile phone?

Alec Glassman: Yeah. So, so right now, and this was some of the major learnings that we got out of all this like really deep customer discovery that we did. What we found is basically there's a lot that needs to change right now to fix the scam epidemic. But for starters, what we really need is we need to give older adults tools to fight back and to be able to access help the second they need it. And so that's what's informed our product design. And so we're available over text, over email, you can chat with us online. And the whole idea is we're this really simple, straightforward, accessible tool where we'll basically coach you through these situations. So you'll reach out to us.

People don't people have been dealing with stuff themselves their entire lives. They don't want to go to their kids. They want to deal with this themsel, right? And so now, instead of going to your kid, what you can do is you just text us, you forward the email to us, and we'll go through, we'll ask you relevant follow-up questions, we'll coach you through that, and then once we have enough information to really discern this, like what's going on in the situation, we'll give you a risk level, we'll give you action items, we'll give you really helpful resources, which people love. Will report to the FBI and the FTC, which people really appreciate, and we'll even help you share it and warn your community. And so, like, in essence, we're turning people from targets to heroes. And that's really what we're doing.

Ryan Eppley: Yeah. Congratulations again on all the, you know, fantastic traction. When you're not in nursing homes and you're here in the city, where's home base for you? Whether you have an office and a team here or your personal home, like where are you here in the city of New York?

Alec Glassman: So, the whole team is based in New York and, I'm over in Chelsea. So, it was a very very easy walk coming over here.

Ryan Eppley: Yeah. Fantastic. Well, where can your camera's right there. Where can our viewers like find you and follow?

Alec Glassman: Yeah. So, silvershield.ai is our website. You can check out our LinkedIn, too. My name's Alec Glassman, so you can find me on LinkedIn. And if you're in New York City, I'm I'm here a lot, so you can probably find me there, too.

Ryan Eppley: All right. Fantastic. Thanks for coming by.

Alec Glassman: Cool. Really appreciate it. Thanks.

Rob May: All right. Our next company is Peyman from we call it Krnel. So, this one's a little more technical than some of the other companies we're talking to, but give us the oneliner on what Krnel does.

Peyman Faratin: We secure AI models at runtime by looking inside the model as opposed to having another model look at the model, right? Security and control, that's what we're building, right?

Rob May: And how does that work? Like, if I if I'm building models or I'm building agents or something like that, how do I interact with your software to do this?

Peyman Faratin: Yeah, our primary users are agent developers and what we provide is a layer of abstraction SDK or some primitives that you can interact with the model internals. So what we do is we look inside the model at runtime as as the model is processing the inputs the prompts it actually creates all these transformations. The T in the G ChatGPT is essentially transformations of these information inside and we hook into the model at inference time and we allow the application layer to see what those primitives are and we do detection and control for the application layer. So you just call the API that we provide to the application.

Ryan Eppley: Incredibly important company that you're building. How did you get into this?

Peyman Faratin: Yeah. I did my PhD in multi-agent systems in 2000. And I have been really interested in these technologies for many years. And I think there is a when we started playing around with these ideas, we noticed that the models have a language internally. So we see English input and output but actually there's a language that these models have and we're trying to decode it. And we saw this at the AlphaGo when they when the Go player from DeepMind was playing against the Korean champ world champion. It made a move that nobody had ever seen before. So they these models have these representations inside them that allows whole new discoveries of phenomenas that we don't know. So that's how I got interested. It's like I believe that this is we make a comparison to penicellin or plastic. So they're they were byproduct of some natural processes and we think there's a lot of value in between input and output and we're trying to capture that value.

Ryan Eppley: Yeah. Fantastic. And where are you today in terms of traction? You know when did you start this and what's the ramp up kind of been like?

Peyman Faratin: The ramp up has been very hard because we have a lot of teaching to do. The market is difficult to metabolize these and finding the right language has been challenging. But are we looking at hardware providers we want to be potentially on the on the wire at the infant time on the hardware itself so you can ship the biases that we have is that you can build this at the application layer but we think the real long-term value is actually being on the wire at every inference time. So these are natural players are hardware providers GPU providers our customers and financial services we're looking at those as well.

Rob May: I mean it makes a lot of sense because you know the things that you're saying there's we talked about this in the last episode there's a there's a AI maturity curve right that people go through and the vast majority of companies in the world are still at like stage one which is like build a proof of concept on a frontier model just to see what works and so but one of the things that we talked about with with Jason Hiner from from Deep View is that going into 2026 and this is going to be a year where people start to think about AI systems and how they really roll things out at scale and all the things around it that they need other than just can the model do the task.

Peyman Faratin: Yes. So I think this is really really really important.

Rob May: How do you guys think about the you know how do you guys think about as where this goes as models get bigger and better and all that like do the do the internal belief systems and representations and structures get more complicated to deal with or do they get simplified because the at scale there's a better intermediate language or what is that like?

Peyman Faratin: Yeah, that's a great question. We don't know but and I think our strategy what we're trying to do is we believe that these complexities one particular lab or entity can't provision alignment teams within these frontier labs are doing great work but there's always going to be incompleteness. There's results from adversarial machine learning that tells us this is the case as well as the theory tells us. So we what we're trying to do is bring these tools to any democratize just in the same way Tim Berners Lee opened up the web by taking HTML and reducing it to HTML. So anybody could look at view source and replicate the internet that's that that's how the internet scaled and what we want to do is bring these vulnerabilities in have many more eyeballs to be able to look at vulnerabilities of the models and I think that's the mission is to democratize these tools. So we don't have to be do doing the discovery ourselves. The essentially the kind of is a whole market of discovery by agent developers themselves.

Rob May: And will this work with models like not just LLMs but if I build specific models at my company if I build SLMs like yes does your tool work with those?

Peyman Faratin: Yes. As we just need access to the inference fabric which is so we look we work with open weight models. Okay. But we have a strategy for closed weight models. So we built we have built applications around Claude Code for example. So as code is running we can essentially we have a digital twin that we can essentially check its internal reasoning and outputs and flag to the application there's vulnerabilities.

Ryan Eppley: How big is your team today?

Peyman Faratin: Two people. Two people located here in the city. Yes. We're on 45th and Lex.

Ryan Eppley: Fantastic. Where that's your camera right there. Where can our viewers follow you and find you internet? Any of the social medias that you might have.

Peyman Faratin: Yes. We are on krnel.ai on the web and we're on LinkedIn and X as well.

Rob May: Fantastic. Well, thanks for coming by the show. We really appreciate it.

Peyman Faratin: Thank you.

Rob May: Yeah. So, our next guest I met when we sat next to each other at a dinner put on by B partners and some other people. But, Sergey from Koop, give us the oneliner about what coupe does.

Sergey Litvinenko: We help companies grow by taking care of their regulatory requirements. We do all kinds of industries, all kinds of requirements, but the key idea is certain to point you'll have to show compliance with those things and it can get nasty and we just take care of that.

Rob May: Yeah. And it's fun to be on the show with you because I'm a customer, you know, you know, you've been helping me with compliance. Expect to see you.

Sergey Litvinenko: Yeah. No, it's very funny coincidence, but yeah, super happy to run into you.

Ryan Eppley: You know, how did you get into this in the first place?

Sergey Litvinenko: Well, I'm going to start with the with the story how it all happened. We originally started in the insurance space in the vision that we had we wanted to protect a new generation of companies from all kinds of risks and we built the insurance company and as we were selling insurance which came with a certain software component we bumped into compliance issues ourselves. For example we started by getting some smaller customers on board and then finally we got a whale. It was like a four billion dollar funded startup which was huge and we went through the whole motion went against the competitors like 100-y old insurance companies and we won based on the value

and I remember that feeling when you know the board meeting was next week and I was like well I'm going to share this I was like this is the first big win and then in between winning the customer on on like verbally to share in this I get this email hey this is like some some easy stuff vendor questionnaire and I open the questionnaire it's holy it's 400 questions all kinds of stuff and one of the things was so and I didn't know about socku and as I researched like three months of work $60,000 I was like well the board is going to think I'm an idiot that I haven't thought about it and I distinctly remember that feeling that feeling partly inspired the current iteration of our company where we said look there are all kinds of things that people don't know about that we can help them take care of that and insurance was the start then we got into the compliance regulatory frameworks and now we do a lot more.

So that was the personal story that personal feeling that led to the creation of our company of the compliance offer in particularly and then off we go.

Rob May: So, and where does some of the AI come into play? Like how do you apply AI to this use case to make it more efficient or better or cheaper or anything like that?

Sergey Litvinenko: There are actually two sides to this. So first, how do we use AI? We use AI to go through your documentation and evidence. So whenever you upload a document or upload a screenshot or the data gets sent over is a JSON file whatever we use AI to understand it and then we can tell you hey is it a good fit is it a bad fit for that particular test for that particular control okay it's also done with a human in a loop to make sure that everything is accurate but that's how we use AI and that saves a lot of time to people like yourself flip that we also do compliance for AI AI. So companies that use AI in their product, in their engineering, or if their AI touches customer data, you need to govern that thing because you don't want to be sending sensitive information to those AI chatbots. So there is this new field called AI governance, which we are taking a stab at and we think is going to be as big as sock 2 if not bigger. And u I think it will be really important for people to make sure that whatever they're doing is with AI is kosher.

Rob May: And you mentioned SOC 2 which is a sort of you know agreed upon standard has some groups behind it. When you talk about AI governance are there standards yet like where is that in the evolution of the sophistication and maturity of the of the market?

Sergey Litvinenko: There are standards I can throw a couple at you like IO the U international safety organization. ISO 4201 pretty much tells you how to manage AI systems but the challenge is that is pretty vague and then there are some companies who are trying to develop like the internal frameworks like I know Meta has something and Microsoft has something but it's only if you sell into Microsoft there are some startups who are trying to do this but so far there's been no universally adopted framework for AI it's going to happen sooner or later there is a right now the government is trying to come up with some framework and the state's trying to come up with a framework. So we'll have to take that into account. I think there's going to be a lot of stuff. We'll probably see a new iteration of sock 2 which talks about AI and we are framework agnostic. We'll work with whatever gets adopted. The most important thing is that we'll make it very easy for you so you can spend more budget on growth not on this stuff.

Ryan Eppley: Fantastic. And how how long you've been doing this now and what's the traction been like?

Sergey Litvinenko: So we've been doing this since 2023. If my memory doesn't fail me, since the seed round, we've grown 23x.

Ryan Eppley: Wow.

Sergey Litvinenko: So, it was a really good journey and obviously we'll have to repeat that a few times to get to the goals that we want to achieve, but the uptake has been really promising.

Ryan Eppley: Absolutely. Where's your office?

Sergey Litvinenko: Jersey City, right across the Hudson River.

Ryan Eppley: So still qualifies.

Sergey Litvinenko: I know, but I'm financially responsible. That's why I'm not in there. So, yeah.

Ryan Eppley: Right. No, we love it. We love it. And your camera's right over there. Where can our followers, you know, find you online?

Sergey Litvinenko: Go to koop.ai. You have lots of stuff there. I'm also on LinkedIn and X, where I follow you. If you type my name, Sergey Litvinenko, you'll find me.

Rob May: Awesome. Well, thanks for coming by the show, man.

Sergey Litvinenko: Thank you guys.

Rob May: Yeah. All right. So, our next guest, Akash Nandi, I met when he worked for a company I invested in. It was a robotics company. So, known for a while. Akos, welcome to the podcast and tell us the oneliner on what what Tendrel does.

Akash Nandi: Yeah, in a nutshell, we fuse all of the horsepower that comes with modern AI with all of the proprietary industrial knowledge that factories and manufacturing environments create every day. And we fuse those to help operators run their job a little bit easier, minimize downtime, maximize production, and give them just slightly sharper reflexes.

Rob May: Does that require a hybrid model? Is it pure software or does that require some hardware and sensors and stuff like that or what do you guys have to do for that?

Akash Nandi: Usually these lines will have an MEES that has sensors, real-time data coming off of it, but that data more often than not ends up in some siloed, you know, repository not actually being used. And I think to really unlock that, being able to combine historical events, past fault records, resolutions with, you know, the power of these LLMs allows operators to preempt, predict, and react just a little bit quicker than they could when before they were just operating constantly on the reflex.

Ryan Eppley: What were you doing, you know, before you got into this? Before startups, before this specific one?

Akash Nandi: Before this one? Yes. So, Rob and I met with my first flavor of startup, which was in the robotics space, automating, farm labor. There's a really gnarly problem tackling the bottleneck of how we feed ourselves, which is getting the hands to harvest the food that we eat every day. We helped, you know, scale that company from a couple founders to an acquisition by a very large food producer in Kentucky. And that was when you I moved down to work with the Inquirer in Lexington, Kentucky in the heart of Appalachia to get my hands dirty. And pretty much the first day I was there, I very quickly realized that robots were only going to be able to do so much. That if you want to unlock these manual industries, you have to force multiply the people who are actually there. And that was my my shift from hardware to software and ultimately founding Tendrel.

Rob May: Fantastic. And so who's a typical customer for you these days? Are you selling to small manufacturers? Are you selling to large manufacturers? Are you selling to the plants themselves? Or are you selling in at like the big company level where they do some, you know, contracts with the plants or what's how that works?

Akash Nandi: Usually, it's the person who feels this pain the most. So, it's at the plant level. It's that, you know, VP of operations, the general manager, the person who when that red light is flashing that the equipment is down, they're the ones who are on the hook. And so we work directly with the boots on the ground, those operators running those plants, keeping the lights on. And typically from a company size, it's usually right between the two ends of the spectrum. So think upper middle market, you nine or 10 figures in top line. So they're big enough to have, you know, the data infrastructure already in place, but they're typically not big enough to have a data science team to utilize all that information.

Ryan Eppley: You're located here in the city.

Akash Nandi: I am.

Ryan Eppley: What part of the city is your team at?

Akash Nandi: So the team is remote. We got some folks in Seattle myself, my product lead and some advisers and investors all in New York. And that was actually a big impetus for me moving back after a 10-year sbatical spanning LA, Boston, Kentucky was a little bit closer to home, but then a little bit closer to the action as well. Fantastic.

Rob May: So, well, I was going to say, so even for something like manufacturing, you feel like New York City puts you close to the action.

Akash Nandi: It puts me close to the airports that get me close to the action. These manufacturers are everywhere. Certainly some on the east coast, west coast, and then, you know, most definitely in the middle of the country. And so, the beauty of this job has been that, you know, when I was living from New York to LA to Boston, I was always flying over what would ultimately be my customers. The best part of, you know, Tendrel is being able to spend a little time in a part of the country that typically hasn't had a ton of investment or innovation. And that's perhaps one of the most interesting things I learned working at at at Rude AI and all of the startup work since is that there's a sense of gratitude and appreciation when folks in these plants see some effort being made to modernize the infrastructure that they've been struggling with for like 30 years.

Ryan Eppley: How big is your team today?

Akash Nandi: We're lean and mean. It's three people, myself, a product manager, and an engineer. We've got some advisers who are subject matter experts in the realms of manufacturing and AI. But, you know, the biggest shift from what I've seen from the first startup today is just how much you can get done with a much leaner, more dynamic team.

Ryan Eppley: Fantastic. Well, your camera's right there. Tell our followers where they can, you know, keep on on your journey.

Akash Nandi: Yeah. The company is Tendrel spelled T N D R E L. Feel free to reach out to me directly accash.io. I'm on LinkedIn and we're looking for design partners and customers in the manufacturing space. So if you are working in a high volume continuous flow environment with really complex equipment where downtime ranges from expensive to catastrophic, drop me a note and we'd be happy to work with you.

Rob May: Oh yeah, thanks for having pleasure.

Akash Nandi: All right, good seeing you, Rob.

Rob May: Yeah, you too. All right, our next guest is Jason Patel from OpenForge. Jason, welcome to the podcast and why don't you give us the oneliner on what OpenForge does?

Jason Patel: Yeah, absolutely. Thank you for having me. OpenForge is a platform of AI agents that learns about your business and then does all the AI SEO work for you.

Ryan Eppley: So, a lot of people familiar with the term SEO. What are we calling it in this new life? So, so like what why do we need this now? I mean, some people say SEO is dying. Some people say they don't like what what what do you see and what do we call this new world?

Jason Patel: SEO is evolving. It's not dying. People who are saying it's dying or they're just trying to sell you yellow journalism there. It's evolving into something much bigger and better. In terms of the name, I've heard GEO, which I personally don't like because GEO is geography. I don't know why you would call it generative engine optimization. Yeah. I like answer engine optimization, AEO, but there's some influencers in the space who think it sounds like EI eio like the rhyme. So, I guess that one's noticed. What I think I call it AI SEO just to make it simple for everyone, but I do believe it's going to be SEO ultimately. Search engine optimization will become search everywhere optimization and that'll be that.

Ryan Eppley: And so, how does this actually work on the back end? Like are you taking docs from the companies, you're scraping their website, getting all the kind of company juice, and then you're like, "All right, this is how we're going to tee it up for all the models to ingest it easiest." Like, how does that really work on the back end?

Jason Patel: Exactly. So, a what our competitors do by and large is that they take a set number of prompts. They make the customer give them prompts and then they give you the data based on the prompts. What we do is that we'll take documents, data, and then also our AI agents will do all the work in crawling the internet for your public facing documents and anything else that's available. It'll then create a tone of voice, a brand voice, everything in your repository of data. And then it'll give you hundreds of queries to choose from. You can add your own if you like. It'll give you citations. It'll give you competitor citations. You can spy on competitors, build new content, refresh content, build third party citations, find the journalists who've written about your competitors. You could do it all on the platform because our AI learns about not just your prompts, but your data and the rest of your market as well.

Rob May: And how much of your AI technology are you able to offload to sort of like standard frontier models? You know, have you had to build stuff internally that's specific to your company? Like where where does the technology fit?

Jason Patel: Yeah. So we have our own proprietary workflow where we use the different models to figure out what works best for certain outputs. So we use we use them all and I think that's the best approach because it's a meritocracy. You want the best model to win. So at the same time we've built our own proprietary workflows on top of those frontier models in order to create the best content that's specific to the company based on what their market is to find different journalist contact information which our AI does as well. And so I think everything is built on top of the foundational layer of the LLM. So OpenAI, Perplexity, they all deserve their their flowers for what they've done for the economy. But then we've built our own tech on top of it.

Ryan Eppley: And how long you been doing this and what's traction been like?

Jason Patel: Yeah, so we launched on officially launched actually we started taking paid customers on August 1st. We haven't officially launched anything yet. We're healthy six digits in revenue and ARR. Doing very well, very profitable. We have a team of seven. We have not taken on any money. We have a lot of inbound investor interest, but we're focused on creating the ultimate customer experience, finding as much as possible about the market, looking at the different changes, and then really just growing based off our customers loving us. Our we have spent zero dollars on hype, zero dollars on paid ads. Everything has been organic and word of mouth, and that's how we know we're on to something. So, August 1, it's been what, four, four months now, and doing very well. It's great. And I know Rob through a different postexit founder community and this is my second company now. I exited their first one. So I made all the dumb mistakes the first time around. So I know all the things not to do, all the things to do to really jumpst start that growth for us.

Ryan Eppley: Fantastic. Yeah. Your team you said has seven people located here in New York with you inside and outside of New York.

Jason Patel: Yes. So we have u my co-founder is on the east coast. He's actually moving to New York very soon. Then we have two developers who are one outside the country, one here. We have a videographer who's international and then we have an operations person who's here with us. So I think that's six or seven. We're hiring someone new, a growth marketer which we can talk about at the end. But ultimately you know we have people within the country outside of the country. Our vision here though is to create an iconic American company not just an iconic Silicon Valley company. And so we want to hire people from the tri-state area from within the country of United States. So we could really you know, just have the DNA of the US within our company. I feel very passionate about that.

Rob May: Oh yeah. I mean, it seems like this with with all the, you know, advertising and media and SEO history of New York. This is a great place to build a company like this.

Jason Patel: I, you know, I was born in New Jersey and so I have a lot of love for the tri-state area. And the big reason I want to build a company here is because there's I'm the son of immigrants and I am indebted to this country for what is it has provided my parents and then for what has provided me as as someone who was born here. And therefore I want to create a company where we're able to imbue different walks of life within the company. You know, not just get a bunch of IV leaguers or not just get a bunch of Silicon Valley people, but get people from all walks of life who know different sales strategies. They come from different communities in order to build again not just an iconic tech company from Silicon Valley but build an iconic American company from you know built by people from from all walks of life.

You know why build in New York? New York is you know it's got a lot of diversity but not not diversity of identity that doesn't matter to me. What matters is walks of life what you've been through different you know types of thought you know thought leadership and things of that nature. You we this city brings together so many people who think differently and if you're able to think differently you bring you know a better repertoire of experience within the company and that's how you grow fast in my opinion.

Ryan Eppley: Yeah. Fantastic. Well, your camera's there. Tell our followers where they can keep up with your journey.

Jason Patel: You could follow me at on YouTube at OpenForge AI. We have 50,000 subscribers. We're the biggest outlet in the space for AI SEO. But you can also find me on Instagram, Jason Patel 13. XJason Patel 13 or of course LinkedIn Jason Patel/Jason Patel. Find me there.

Rob May: Awesome. Well, thanks for coming on the show, man.

Jason Patel: Thank you very much for having me. Great work, guys.

Rob May: Thank you. Okay, so those are our startups today. Definitely if any of those you saw were interesting, please go check them out and support the AI community here in New York. The more that we can help these companies grow, the more customers we can bring them, the more that the AI community here is going to become and remain one of the best AI hubs in the world. So, as we wrap up 2025, I mean, it's a shame Anna's not here today. She already took off for the holidays. We still got to get Annabot up and running.

Ryan Eppley: We got to get Annabot up and running. Thoughts or reflections on on the first couple of months of the show here? You know, I felt like the amount of talent we had on, you know, sitting in these chairs was was really fantastic. I think just everyone has different perspectives about what does AI mean for them. And as you've always said, I feel like everyone here is trying to apply AI in a really what's it doing for me today? What value is it going to create for, you know, my customers, my team, you know, today rather than what's what's the future in five years, you know, which is something that I'm really proud about here in New York in the applied AI capital of the world.

Rob May: Yeah, I think a lot less hype and a lot more you know, realistic progress than than maybe you see in other geos.

Ryan Eppley: Yeah. Looking at you. And thanks for thanks for all the followers that have helped built this up. You know, we really appreciate all the comments, all the subscriptions, and all the help. You know, when Rob first told me he's he's trying to put together AI in New York City, you know, I said, "Really, Rob, another podcast? Does does the world need that?" And honestly, it's been really fantastic getting everyone's, you know, input on on how much it helps them grow their own company, get, you know, the latest on on the news and, you know, try to keep it real and raw.

Rob May: Yeah. So, you'll see some changes going into 2026. We've learned a lot from these early episodes. So, we're going to mix some things up, try some new some new ideas, new formats, hopefully have some sponsors, knock on wood, and probably have a AI and NYC conference sometime in 2026. So, thank you guys for watching for for all your support. Please go watch the other episodes, give us feedback, like us, subscribe, all the things you're supposed to do, but most importantly, send us ideas, send us guests. We hope you have a happy holiday and a happy new year, and we'll see you in 2026.

Heat. Heat.