# How we did it: True Horizon on scaling a business with n8n

## Метаданные

- **Канал:** n8n
- **YouTube:** https://www.youtube.com/watch?v=W25ECU7bMuU
- **Дата:** 23.02.2026
- **Длительность:** 24:31
- **Просмотры:** 1,901

## Описание

n8n sits down with True Horizon co-founders Tyler Baughcome & Milan Tahliani to discuss all things automation. Having already grown their business impressively in little over a year, we hear about the past, present and future of True Horizon, as well as their predictions for automation in general.

## Содержание

### [0:00](https://www.youtube.com/watch?v=W25ECU7bMuU) Segment 1 (00:00 - 05:00)

My name is Milan. I'm co-founder and CEO here at True Horizon. — My name is Tyler, co-founder and CTO of True Horizon. — So, True Horizon is an AI consultancy. We specialize in leveraging NADN to drive returns for our clients by implementing custom AI solutions, AI agents, and AI strategy. Yeah. So we met uh we were both working as software engineers at a company where we were um interns there. Actually that's when we developed a pretty close relationship. Realized that we both have ambitions to uh leverage technology but also start our own company. Uh fast forward about a year and a half later, it's now fall of 2024 and AI and automation is just going crazy. Nad is just beginning to catch fire. there's a new foundational model every single day and I basically phone up Tyler and say, "Uh, hey man, want to start a company together? " — Yeah. No, I was um at that time I was doing my masters in quantitative finance. Um I was not enjoying it because it's quantitative finance. Um and I get a call from Milan. He's like, "Hey man, want to just start a company? " Like it was very impromptu. Um and so I left my degree. We started doing some small things on the side and then we uh we met Nate. — Yeah. So, in December of that same year, we partnered up with Nate Herk, who's our third co-founder, uh, and officially got together and, uh, formed the band, so to speak, in January of 2025. Uh, so we've been in business for about a year. We've done over 25 engagements, shipped over 100 AI projects at this point. Uh, we've got a team of about 15 incredible engineers, and are looking to continue scaling. — I have a computer science degree which at its core is automation. Computer science is found on the basis of automating processes and making them faster anyway. So technically I've been doing that for 5 years or six years now whenever I started my degree. Um but then actually framing it in terms of like it's the word automation and AI and you would do it in low code that probably started for me um honestly when when Milan called me about a year and a half ago. Um but the transition was sufficiently simple like I had been building things for like he said internships where um we were automating tax processes um and I worked at a fence firm where we were automating certain elements of well actually I can't say but um so I would say the journey started a long time ago but within low code platforms like inad a year and a half. Yeah. — Yeah. I think for me it was um I actually remember it very distinctly is right when Chad Gubt came out I was working on a computer science problem set. I also have a computer science degree. So the homework assignments are problem sets. And uh I remember my buddy came running and knocking on my door. He's like Mill and Millan you got to see this. It was absolutely insane. He copied and pasted our homework assignment into chat GBT. Boom. Started spitting out, you know, hundreds of lines of code. And of course at first plug and play it didn't work. But I think right there I saw the potential. Uh initially I was like hey I can you know automate my homework. Uh and that was pretty incredible and then you know as time progressed realized you want to build more and more complex systems realized you need a tool like nan that makes it a lot easier with a graphical user interface by which you can actually automate processes and start to build some really powerful stuff. So you know how we started versus where we are now it's radically different. When we formed the company, again, it was us three co-founders and we were in the trenches doing literally everything. So, if you think about the life cycle or the value chain of an AI deal or an AI consultancy, it starts with the lead generation. So, where do you get the prospect? How do you sell them and convince them to invest money in you and your company? Um, doing all the contracting, NDAs, MSAs, and whatnot. And then actually fulfilling on the work, developing and deploying and delivering an AI agent or an LLM system. And so in the beginning, all three of us were basically doing every part of that process where it was literally from the beginning where a prospect comes in, Nate, me, and sometimes Tyler would be on a sales call, we sell them. At times we were all working on a contract, and then came the delivery and fulfillment. There was a time where even I was also in NAN or inside our IDE and actually programming the solution. And uh yeah, you know, Tyler's laughing at me because nowadays I'm not allowed anywhere near our production solutions. And um it's evolved quite a bit. bit and we've all I think had to kind of grow up and step into our own role as leaders of our own respective portion of the organization and uh yeah, it's changed quite a bit. Yeah, I would say the place where we've scaled on the personnel side is mostly technically like Milan still operates the majority of the sales um and the executive operations but most of our

### [5:00](https://www.youtube.com/watch?v=W25ECU7bMuU&t=300s) Segment 2 (05:00 - 10:00)

team are engineers. So at the beginning it was myself and one other person that we had brought on board. So we had a team of two delivering at the time 10 to 15 solutions. Um so around the January March mark we were just bombarded and I was probably spending 16 hours a day just delivering on the solutions personally. Um and over time we brought on more and sort of scaled down the number of clients so that it made sense so none of us were suffering. Um and that numbers just progressed so there's more of a hierarchy in place on the technical side. I have a VP of engineering. Um and then we have [snorts] I guess you could say project managers and then a subset of engineers on individual projects or departments. Um, so we went from, you know, no structure and chaos to to more organized. We're certainly not, you know, the most structured, but definitely better now. — Yeah. And I will say the way we operate also has evolved significantly. Like in the beginning, in those early days, we signed 10 to 15 clients all at one time, just a three-man team. That was brutal. you know, we were all just sleepless nights over and over again, making sure the client's happy, making sure the project's on time. And um it really comes down to the way we were operating. You know, we were essentially positioning ourselves as a dev shop, you know, just a development partner where we don't really do anything except code and build solutions. But the reality is the exact opposite. We are a holistic full-on AI partner. And we realized back then what we called our pre-sales process was actually a full-on consulting and strategy and scoping phase. And so now we've shifted the way we operate. We've shifted our business model to be more so a holistic AI strategy partner rather than just a development shop. And turns out that's what 99% of the market needs. Anyway, you know, at one point we had 15 projects going on simultaneously with a pretty small team. And that's when we realized like, hey, we need to pump the brakes and really figure out what it is that we're doing here. So, we decided to pause taking on new clients, ship and deliver everything we were already contracted to the best of our abilities. And in that time, really refocus our strategy, refocus how we're positioning ourselves, and re and refocus the value that we actually deliver to clients. And that's where the shift came from development shop to holistic AI partner. — Yeah. Yeah, there was never a time where we didn't have a client, but it went from 15 to one or two. Um, and yeah, we just need more structure at that time, I think, on the technical delivery side of things. Um, we had like canband boards and some amount of QA procedures, but nothing rigid and nothing standardized. — So my first encounter with NAND was I was developing a voice agent and shipping uh some projects for some clients of mine and previously I had gone you know fully bespoke custom code and it just so happened I was just on YouTube and I came across Naden as a platform. [clears throat] I was like, "Wow, this looks pretty cool. " You know, built on top of lang chain. Seems like it's pretty powerful, pretty flexible. And that was basically how I dove head first into the world of no code, low code. Saw that there was other platforms out there. Um, and basically began a full evaluation process as to which one would suit my needs the best. So was building out the same project on a bunch of platforms you know naden relevance make zapier and uh came to the conclusion pretty quickly that nadn was the way to go. Yeah. So my first encounter in Aaden was in the pre-truth horizon era. We were prototyping something for a possible discovery call and I had been building it in two other platforms which I will not name but I actually was having an okay time with those and somebody was like you know actually you should build it in inate because here's a template or whatever and I didn't like it at first. It felt foreign to me. Um but you know it's certainly evolved since then and I would say I wouldn't say I'm in love with the product. I would say it's weird to fall in love with the software, but I you know it's there's a reason we continue to use it every day and uh why you know I asked that my engineers use it continually. Yeah. But my first encounter wasn't uh in love at first sight you know we had to work it out in our relationship me and Naden. Yeah. — Yeah. And one of the critical things too about NAND was the community. Like I distinctly remember running into a bug on both NAND and one of its competitors. this case it was relevant AI and uh to debug the same problem. I was having effectively the same problem on each platform. Looked it up in any tons of community forums, threads, people running into the same problem and a massive community effort that felt like it was going to help me like wow here are these hundreds of people that want to help people like me. Same thing on the other platform simply just did not exist. And so I think that was a massive compelling reason as well.

### [10:00](https://www.youtube.com/watch?v=W25ECU7bMuU&t=600s) Segment 3 (10:00 - 15:00)

— Yeah. The useful aspects of the community I would say communities like Nate for instance are really special in the sense of the ambition you see in terms of the innate community itself like obviously uh a gold mine of debugging problems. I mean I've probably been on if I counted thousands of forms by now just checking how to solve a problem. So in that sense it's very unusual but in does seem to have as softwares go and I've used many softwares as a computer scientist you don't really encounter a form where people are like happy that they're continuously using something and like I get that. So I would definitely say like you know I can resonate with the positivity that strangely lives around Naden probably more appropriately answered how do we use it minuteby minute because it is the core of our company um in terms of clients I mean every single solution we deliver thus far has been well I would say 97% of what we delivered is in Aaden with some caveats um so on a day-to-day basis I mean client wants a solution to be delivered. Um, and of course there's an occasion where client demands that it's in because they like the product and you know it's a happy marriage between us and them. But I've always maintained to my engineers and myself and my team that we build however we need to in order to get the job done and so happens that in end is the choice that we have made. Um, so obviously customer has requirements we start building. There's probably a call every single day where we show to the client or the stakeholder what we have built with a small demo and it is nice enough where you can show that and it actually makes sense to somebody. It's not just lines of code executing. Um in terms of I think we personally use it as well. I mean I've built you know dozens of small automations to make things just easier for me. For instance, we live in client environments with client laptops and I have to sync calendars and I've built the automations that actually do that. unfortunately um and small things like that just on a recurring basis where you know it can be done more efficient and so you just choose to build the thing that can yeah — yeah I think uh yeah you know the way we use NAN like Tyler said it's probably better counted minute by minute or asked and how do we not use NAN because it is you know like Tyler was mentioning core to every single one of our internal processes the projects that we ship for clients so we've really been able to build a fantastic business around this product going way beyond just back office or backend automations, you know, being able to connect these amazing backend functions, these LLM and AI capabilities that are getting smarter and better every day with bespoke frontends. And in fact, we even run the frontends in NN now. Tyler has like a nifty way to send a web hook that has HTML and we're spinning up frontends uh in NN now as well. So, it's a full stack tool. It's how we operate a lean machine and um it's how we're planning on continuing to scale into uh 2026. — Yeah, there's not a minute in the day where someone on our team is not using it in probably given the time zones that we have because we have an Australian — in our midst. So there is probably not a single minute in the day where someone's not in it. — Yeah, that's a great point. We have an engineer in Australia and then moving you know more west I guess we have an engineer in South Africa we've got someone in the UK someone in Egypt and Morocco then you know across the Atlantic obviously we're on the east coast and then we've got someone on Pacific time so quite literally at probably every minute of every single day regardless of where you are someone on the true Horizon team is in NAN So yeah, so one of our key engagements is with a large tax advisory software company. So they have about 6,000 employees, give or take. And uh Tyler whipped up a video series so that we could empower their entire staff to become citizen developers. So folks can utilize Naden within their own life to automate you know maybe the repetitive and mundane parts of their work. Uh that slowly expanded into a department level engagement that then led to a entire seuite executive training where we were fortunate enough to run a 1-hour session which uh included a live edit walkthrough. So you can think about you know it's a10 billion dollar company this powerhouse seuite and they're all following along and building out a personal assistant agent in n that we ended up getting really positive feedback on and many of the seuite enjoyed um that then led to our broader consulting and development engagement where we now forward deploy AI captains or end consultants into departments so they get embedded amongst the stakeholders amongst the workflows and processes and are responsible for leading the AI charge inside of that department not only on the strategy side but also on the implementation and

### [15:00](https://www.youtube.com/watch?v=W25ECU7bMuU&t=900s) Segment 4 (15:00 - 20:00)

driving actual results and I think uh currently we've got probably over 25 active initiatives going on again across departments so use cases within their sales department within their product department within their finance department within their security department some really cool and interesting AI projects going on that uh we think are going to drive a ton of value and really position Avalara as that AI first — You know, what we've seen with the Avalar engagement, what we've seen across our other 25 engagements has been absolutely incredible. I mean, stunning ROI numbers. And, you know, there's a lot of hype in AI right now. There's a potential AI bubble. There's a potential overinvestment of VC into products. I think all of that could be true and there's a lot of merit to that. However, that doesn't change the fact that organizations who can adopt NADN and leverage AI are seeing incredible gains. You know, whether or not VC is overleveraged is irrespective of the fact that AI consultants can come in and make a massive impact inside your organization. So, I'd love to see Naden driving and taking those enterprise engagements kind of to the next level. Um, because I think the returns are absolutely there. Um, I'd also love to see NN continuing a partnership with True Horizon. know that that would be absolutely fantastic and uh investing in the partner network with events like these are absolutely incredible. They achieve a level of buyin from partners such as ourselves that is not otherwise um it's not otherwise attainable. — The game show from Max as well is very interesting. I heard a lot of ambitions of like we want to do this for the community and then one happened and I think that's I think that is a really cool concept you know incentivizing people to learn it. Um but yeah, I just you know I just want to see NAN grow. I think in particular NAN has a lot of attention in let's say more casual settings but I think NAN is something that educational institutions need to know about as well. I have a computer science degree and we learned a variety of languages but you know we're in a space now where people are writing code for their homework and the like and I think the paradigm shift should be something like instead of generating the code that answers your questions let's actually embed AI into the very material that we're teaching which is enabled by something like NAN and I think if college students were coming out with like I'm proficient in something like NAN you know that's I would see that as a win for not only in but overall AI enablement of you know developers. — Yeah, I think I'd love to see NADN really go head first into its own AI features. So it's obviously a low code no code builder. You could build incredibly powerful AI agents on the platform. I'd love to see NAD have more of an AI first component. So you know being able to come and just speak a workflow into existence and it spins up on NAN and it debugs itself. And I know that's the direction that uh that NADN is headed in and I'd love to see that become a kind of core part of what it is that NIDAN is doing. I think the way that we're interacting with AI as it stands today is somewhat outdated. You know, I think a lot of people are there on a chat GBT interface and they type in, they get an output, then they copy and paste that output into email and then they're manually checking. I think that in and of itself kind of eliminates the value prop of AI. you know, if you have to copy and paste output directly from a chat interface. And so, I'd love to see NAN kind of overcome that barrier of user experience. And I think one of the ways that they can do that is with a voice to workflow creator that just works. So, what do I see for the automation space moving forward and into the future? It's very interesting and it's kind of a abstract slash neverbeforered question, right? Never before have we seen technology that can recursively self-improve continue to get better every single day. And I think the answer to that question is going to lie in, you know, basically what I just said, right, which is essentially the intelligence of the foundational model. How much smarter and how much faster are the foundational models going to continue to learn and eventually hit that singularity where they just have exponential recursive self-improvement because you know when we get to that point we're now discussing super intelligence that goes far beyond human intelligence right I mean it quite literally thinks in dimensions that humans cannot conceive uh computationally that's you know what's happening when an LLM is computing its answer Right. And so I think for automation what that means is potentially the foundational model is going to be so smart that it can just go ahead and do things on its own. It can write its own automations. It can

### [20:00](https://www.youtube.com/watch?v=W25ECU7bMuU&t=1200s) Segment 5 (20:00 - 24:00)

execute tell you where are the most optimal places for automation. And so I think the answer to the question really lies in how fast and how much smarter are the foundational models going to continue to get every single day. And from what we're seeing it's uh it's an exponential growth curve. So, it's very exciting, somewhat scary, somewhat doomsdayy, but uh I tend to be more of an optimist and uh and um I think humanity is going to figure it out hopefully. So, yeah. Where I'm interested to see where automation goes is we've always automated in the digital space because data is easy to play with and we control that environment, but we don't control the physical realm or we have a limited understanding of it. So at what point will automation cross into the physical barrier and how easy will it be? We have robots that are you know relatively functional at this stage. I think there's instances where it's impressive. Like I know um this company called Engine AI released a bot that was doing karate with their CEO and then there's another company who are selling a house assistant but when it goes in your house it like I saw a video where I like walked into a mirror and the mirror broke and then it like was burning something in the oven and it's just like we're not quite there in the physical realm yet but like what if you know like I could tell this chair to move like two steps forward or like you know transform a building in some coherent way. So I think the day that automation starts to enter the physical realm in a coherent way is going to be a very dangerous day. But I don't you know there's so many manual tasks that we do that can't aren't in the phys the digital realm like building a house. You know we have construction contractors that come out they lay bricks over bricks and it's like that is a process that if it was in the digital realm we would be automating but because it's physical we don't. So that aspect I you know I think that is the future of automation — and I think it starts in the mundane like hey every time I sign a new client I spin up a Google drive and manually send them my contract and manually wait for the response like I think it starts with that right and then it goes to the next level of you know automating entire workflows entire processes there's all this talk of folks being able to build one man billion dollar companies cuz AI agents and automation are, you know, running the scenes, then I think you get into the physical world. — You almost can't avoid using it. Even no matter what you are, if you're a mom and pop pizza shop, what are you automating today? And how does that help you? Um, and I think, you know, not only can companies not avoid it, but it's just so helpful that you shouldn't, you know, my I my uh my girlfriend and I, we have like a kind of dream for like a dog park business, you know, like we own a space and it's like even that company, right? It's like what are we automating? Are we automating outbound emails? How are we automating like invoice processing and things like that? And it's like doesn't matter what you are, you will automate. Question is just how. Oh, and I guess on the personal front uh I've done a couple interesting things um because you know I I've become proficient in it and over the years and when I turned to my life I'm like what's something interesting I can do? I remember for the club World Cup final, which was in the United States, I had some friends who were like, "Let's go. " And I was like, "Yes, but it's so expensive. " So, I built um I just built a price tracker for tickets. And I would just get notifications like, "Oh, it's, you know, at an all-time low or it's went under $200. " Um I never got a ticket, but — the system worked, — but the system was functional. Yes. — Yeah. I think one of my one of the most creative use cases for automation that I've seen so far is was actually done by one of our partners Brody Automates and he built out what's known as the Hell yeah automation or you know I can call it he's built out what's known as the hecky automation go which is basically every single morning at 6:00 a. m. your phone is going to automatically start playing AC/DC. So you know you'll wake up to some heavy rock. It's going to send a customized motiv motivational message based on what's on your calendar for that day. but it can also send it to contacts of your choosing so you know you can spread the love and spread the motivation and um and that's a super long automation like it keeps on going but I think just the initial like you know I want to automatically wake up to AC/DC every single day is probably one of the more creative and fun things that I've seen so —

---
*Источник: https://ekstraktznaniy.ru/video/15115*