The data exists. The intelligence doesn't.

All right. So welcome to the corridor. My name is Angel Leon, your host. And right now I think almost everyone building AI is pointing it at the at the easy industries and the ones already sort of like living in the cloud, right? My guest went the other direction into manufacturing, one of the oldest and most stubborn corners of the economy.

where the most valuable information in the in that building is not necessarily in a database. it's in the head right of the operator who has run that machine for 20, 30 years, right? Carlos Maiguel he's the founder and CEO of Corello. and basically an AI native manufacturing intelligence system for small and mid-sized manufacturers.

their sorry, their line is one of the sharpest I have heard so far, honestly, in this in this industry. The data exists and the intelligence does not. So Corello captures what they call tribal knowledge, which I love that terminology, across you know these legacy systems, right? These shops that already run on, right? and it turns it into something the whole team can actually use.

he's a serial founder and he's building this out of Boston. So Carlos, welcome to the corridor.

No, this is great. Thanks. Thanks for having me, Angel. really excited and looking forward to talking with you today.

All right. So let's get let's get started with I guess with the why manufacturing first, right? So of every place you can point AI right now, you chose manufacturing, again, one of the most stubborn and oldest and least glamorous, I would say, right? And sort of like boring industries, right, there is out there. And not many people, you know, want to go that direction, right? So why did you go there? What did you see that made you say

This is the one.

Yeah, no, and maybe a little bit of context about myself because I think it's relevant for the question that you ask. so I'm originally from Colombia. and then I came to the States because before I brought my previous startup to the States in 2019, but of course COVID killed everything. So I jumped into business school here in Boston, got my MBA degree and trying to figure things out what was gonna be my next like venture or adventure, let's call it. I started doing consulting for as for

For specifically small, medium-sized manufacturing companies. And the first day that I went to one of my clients, like going to the customer, I was like really excited. I was like, you know, man, consulting in the manufacturing space in the United States, that's gonna be really awesome. Everything's gonna be state of the art, everything's gonna be organized, everything's gonna be full of robots, automations. And first thing that happens is that my customers facility.

Okay.

Carlos (03:09.75) It's like going all the way back to 1920s. Everything is old, everything is paper-based. I remember they had this 30-year operator who knows everything about the company itself. And then when that person wasn't around, they didn't know what to do with the factory or the operation. So it was like kind of hijacked by the operations manager at the same time. And I was like, well, maybe this is just one in a million. Well, what happened is that

By doing consulting for the past five years, I was like, every company that I vested had exactly the same problem. And it's all the industry, as you mentioned, it's not sexy or it's no one really wants to work in that industry. So there is a huge gap between huge manufacturing companies and small manufacturing companies that are critical path for the manufacturing supply chain of the country. So my customers right now.

They serve huge companies like Lockheed Martin, BAA systems, Electric Boat, companies that are critical for the country itself in many industries. So what's happening right now is that the gap is getting wider and wider because big companies, they can definitely adopt this type of technology. But then the old companies and small companies are left behind. And I've I've seen companies that it's like they're running their businesses in pin.

and paper steel and they do very complex things for the manufacturing supply chain.

So I wanna I wanna sit on a line, right? from your site, specifically, right? because I think it's sort of like the whole thesis, in basically in four words. You know, data exists, intelligence does not, right? unpack that for someone who has never set foot on a factory floor before, right? what is actually happening inside these shops, right? and

And how did you come up with, you know, the thesis itself as in, you know, what is it that you saw that made you think, okay, the data exists, where does it exist, right? And what do you mean by intelligence does not?

Carlos (05:23.82) Yeah, so I think to your question, working like and leaning the trenches with the small manufacturers, I realized that data was everywhere. and really everywhere. So many if you think about manufacturing, you get an input and then you get an output through a whole process. And then how you inform that process, well, that's where the challenge is. every customer that I visit, I ask, okay, so where are you storing data?

Do you have a CRM in place? Do you have an ERP in place? Some of them were like, Well, yeah, we have an ERP, but we don't really use it. We just use like five percent of that solution. Other were like, hey, we store everything in an Excel file. And then when I get access to those files or those silos, I was like, Well, you are pretty much lying to me when you say that you don't have data. Actually, you have tons of data. The thing is that it's so fragmented that you're not able to use it in a smart way.

So what we're trying to do is to, in simple words, is to try to unlock the tribal knowledge that's manifested in different ways and the tribal knowledge that's also trapped in the operator's head. So that now manufacturing companies can have eyes in their operations. Whether you are running your company in an Excel file, I've seen those in what they call paper travelers, which is crazy because it's a paper that travels with the order.

They call it paper traveler or paper router. it describes all the process of the manufacturing company. And you can in those papers they are capturing temperature, the operators, time, quality checks. so all of those scattered data across multiple silos, we are able to ingest. And then what we're doing is we are first giving them ice in the in the operations, but at the same time.

We're giving them specifically ROI on the t while using the technology that do that we have so that they can justify the implementation of these AI native layer of intelligence that we are giving them. because I think this technology is not good to automate things in those places because they are really complex, but it's really great to empower people. And I'm a true advocate of that because I've worked with the people of the floor.

Often they only speak Spanish, Portuguese or French. So I'm speaking Spanish with these people. I'm like speaking in their native language, understanding their challenges, their problems, and understanding the problem that they have and seeing how they are reacting to my technology. It's really fulfilling for me because it's actually helping the people that are building the physical world for us. so that's like another thing that we're working on.

That's really cool. Definitely. Yeah, that's really cool. And to go to the heart of it, right? and coming back to that tribal knowledge, what does that actually look like in practice? You know, and why do you think that no one has solved it before now?

Carlos (08:29.74) Yeah, and I think that there are many, many, many causes of why they haven't solved it right now, or they think they might. but in reality they haven't. so first thing is the industry itself, as we've been telling, it's like high mix, low volume, small manufacturing companies are first of all, really complex environments. So imagine everything changes every day.

So it's not like an assembly line where you get the same input and the same output. Well, that you can actually automate really easy. But then when you have different orders, different products, and different processes to run, well, that's where manufacturing lines get really complex. I think that's the first thing. Secondly, is that s most of these companies are family owned and second, third generation run. So

They already know how to do things. And then if you talk about the things that they do with them, it's like for them, it's really easy or really, let's call it normal, regular, however you want to call it. But they do really complex stuff. Now, when you try to adopt innovation in these places, you need to really show them why is the why this matters to them. And I think that some of the solutions that are out there, like systems of records that are not similar to what we do, but they're still chasing the same dollars.

Are not really showing that ROI for them. And then on the other side is implementation. I have a customer that has been trying to implement an ERP system for the past two years or so and still getting negative ROI. they've spent almost a quarter of a million dollars just in trying to implement that thing and it's already outdated. So being fast, allowing them to understand the ROI is key, and that's something that we really understand.

I think often what happens is that huge software solutions are going to those places, showing a specific ROI, but then when they're out, it's like no one really knows why that matters. So that's why we think that Corello is important because we live within the manufacturing environment.

Yeah, so take us into the product then a little bit, right? I'm a small to mid-size manufacturer and I bring Corello and you know what changes for me Monday morning when I come back, right? And I know you guys have like instructions and quoting and onboarding. So walk us through what the system actually does for that shop.

Carlos (11:03.34) Yes, definitely. And so maybe I'll I'll walk you step by step. let's imagine Angel, you're my new customer. So first thing is that we're gonna start ingesting the data that's useful for you. Within minutes, we are able to ingest data, whether it's structured data, things that come in a data lake, a database, unstructured data, everything that's Excel files, PDF, Word documents that we can also ingest.

Paper-based data with our OCR computer vision technology, we're able to ingest that data and pretty much give you an intelligent way to interact with it. And then the other thing is the tribal knowledge. We are ingesting tribal knowledge by you interacting with Corello and contextualizing all the data that you have with that interaction. So that's the first thing. We ingest your data. Secondly, and that's taking us around some days, we are gonna deploy an AI coworker.

That's gonna be embedded in one of your workflows. So when I'm talking to a customer, we try to prioritize the pain that they the biggest pain that they currently have. Right now, my first customers, and I think this is a common denominator in this space, is quotation. So small manufacturing companies are taking around 10 days to get back to the customer with a quote. With Corello, we were able to ingest their operational data, financial data.

The data exists and the intelligence does not. You have tons of data. The thing is that it's so fragmented that you're not able to use it in a smart way.

and tribal knowledge. We embedded in their email address, not giving them another UI or another dashboard. We think we think that user interface is tending to disappear. So now we embedding their workflow. We embedded what we call Sophia. We give we give the coworker a name and now they are interacting with Sophia through email and they're getting back to the customer in two minutes. That's solving for speed. That's the first thing that

-huh.

Angel León (12:40.23) Was

Wow. That's amazing.

Carlos (12:58.21) Solving. Now they're in they are starting to interact with Sophia. Now Sophia is getting intel more intelligent than before. And they are in they are starting to understand winning rates, starting to increase them. They are optimizing price because if you're working with a huge corporation, well, you can actually charge a little bit more and they'll still take your order. But then when you're working with smaller suppliers, when well you wanna increase the volume of the order. So that's already happening. And

The coolest thing about this is that they all they are already seeing impact in their in their profit margins. That's the whole end goal. So we embed it in their workflow. So now that person comes to the office and instead of going back and forth with files and colleagues, now it's having an intelligent conversation with Sophia that's getting him or her the possibility to do the work faster.

And now that person in specific is pretty much using the spur time to get more customers. That's and that's a common denominator. We have multiple use cases on our on our AI coworkers. The other one that's pretty common is about work instructions. And remember, this is a multimodal AI coworker, so we don't have a UI for everything. Of course, we have our it's called a quote unquote platform that allows people to ingest data, talk to the data.

Got it.

Carlos (14:24.45) build their brain because that's the end goal is that they can build a brain with their data and interact with that brain. Not only interact, but deploy multiple AI coworkers that are giving tangible ROIs for them. The other use case that I was going mention was work instructions. So going back to the example of the of the person that only speaks Spanish, now that person is able to speak in their native language with a document that's written in English. That by itself

It's already given it's already adding capacity to the plan because that person when doesn't know what's happening, that person needed to go and ask the supervisor what's going on with that specific order. Now is having an intelligent conversation with the work instruction. That's fine. But what we are doing is that we are mapping that interaction so that the new work instruction comes with the with the advice or the

the new knowledge of the operator itself. And that it's a it's a constant feedback loop in the data. that's another one that we have in the in the market.

That's very cool. And I love the, you know, the like I think one of the main things that attracts my attention the most, right? It's that and I feel that this is also one of the main things that I'll builder audience, right, wants to hear. what does it take actually to build something right modern like you have and AI native on top of systems that are

decades old or even sometimes even more, right? And where does the hard part really live? You know, is it is it in the technology or is it getting a shop floor to basically trust it? You know what I mean?

Yeah, I think it's it's maybe a couple of things. The first the first thing will be like we are not trying to integrate with their current systems, but what we discovered is that we are using their systems to get data out of them. So as I was saying, if they have an ERP system in place, well we can suck data from that ERP system. If they have an Excel file, we can e extract data from that Excel file. So we are

playing the game of adapting to their production and they don't need to adapt to our way of thinking. That's key because that will that's shortening sales cycles and at the same time implementation time. so that's the game that we're tr that we are playing. that I think that

Definitely. And it works it trains the probably right the employees and all that in a much more like traditional manner that they're used to, even though they're being trained for you know for an AI native right system. And how do how do they interact with Corello, with Sophia? Is it like a voice? Is it like with them on you know typing something? How is that interaction?

So right now with Sophia, depends on the workflow, but the workflow that I mentioned about quotation is strictly through email. So it's like if you were emailing another coworker, but you are emailing your AI coworker, then in the future, and that's also part of the product roadmap, we envision Corello as a multimodal solution in the in the way that we're gonna start capturing data through video, smart classes if you are in the in the front line.

voice if you don't have a screen. Some of the of some of these places they only have a station with a computer. So that chat based interface, that's fine. If you're the CEO of the company, you interact with your data really different than if you are the guy who does sales for the company. So those are things that we are part of the of the product roadmap right now.

Cool. And then your website talks about you know, the AI coworkers and coming back to that again, right? becoming basically the goal driven members of a team, right? And I feel that's sort of a bigger claim than a software tool, right? So where do you actually think this goes? and what does the next few years look like for AI on the factory floor? And what do you think most people building AI right now are getting wrong?

Yeah, so I think we think about our AI coworker as another employee. And in fact, we're trying to map the ROI that we're bringing to these places in different ways, but we want that AI coworker to bring the same value as a new as a as a human coworker. So with metrics, with goals, with context, and with memory, the AI coworker can actually bring value to the company as a human being. So

We are thinking of these coworkers are as you mentioned, goal driven, that they have multiple touches with different people. So if you are the CEO of the company, you interact with the same the same co-worker in a different way than if you are in the floor asking, not asking questions, but generating value totally different than if you are the operations manager. So we think it's goal driven, it's used to increase revenue, it's good to increase margin, it's good to lower cost.

instead of having a specific automation for small tasks. We think this is the bigger vision of what we are building here. And then if you if you think about the manufacturing space in specific, we are thinking about decentralizing the whole manufacturing supply chain of the country by embedding instances of Corello in multiple places so that in the future these coworkers can interact outside of the factory.

but with their suppliers, with their customers, and with their with all the supply chain of this industry in specific.

Amazing. And you started more than one company, right? Like I mentioned at the very beginning, in your path from Colombia to building this in Boston now. So what has this particular journey right taught you that the early ones didn't, right? And what has been the hardest part for you getting Corello off the ground here in the United States?

Yeah, I think being a serial entrepreneur, like that's I think it's important, that this is not my first one. because I mean you get you get building a company is not easy. It's a really lonely place to be, I guess. and through the through the learnings of other of my other startups, I've learned how to pretty much understand the market, how to talk to investors, how to talk to customers, and how to

understand the technology that needs to have a product market fit. So those are like the things that I've that I've reused from my previous experiences. I think biggest thing is having a market and knowing that the product has a fit in it. That's like the biggest thing that you need to have. Because I mean some founders they are chasing for multiple things at the same time. We are really laser focused on our market

High mix, low volume, small manufacturing companies. I know this technology can have multiple applications, but we want to be laser focused at first and then scale from there. so I've been I've I've been using that. And then I think that it's always great to build with more people and more entities. So for example, Corello itself, it's being backed by Scale Up Labs. That's a venture studio in Boston who is building part of the technology of the company.

because you have access to different resources, technical resources. partnered with one of my co-founders, he's the former CTO of Rockwell Automation. My technical co-founder, he brought his first AI startup to the States in 2001. so that's great to build. I prefer to build collectively and not just individually, because I think you can actually get input from many people. And I just got out of the acceleration program from C Ten Labs.

that's a spin out of MIT. So building collectively and openly, that's how I like to do things. So you can get feedback from everywhere and start well, if you fail, fail fast, but people faster.

Yeah, definitely. And thank you, Carlos for all this and taking the time today to speak a little bit more about Corello and your journey, right? I think it's really important for a lot of Latino founders out there to listen, specifically, you know, from people right that have done it before. And like you say, there's some sort of, you know, advantage, right, in that sense because there are things that they

on your first company maybe you actually struggle with that you didn't struggle with Corello, right? and things like that, because you already had that understanding, right? And like you said, how to talk to investors is one of the I think one of the main ones right now for, you know, very early stage Latino founders, understanding fundraising, et cetera, what the process, right? How that works and what to use it for, I think is one of the most important things, right?

what to use the capital for and if you need the capital, right? and when you need the capital. so I appreciate, you know, your story and I love Corello. I think you guys are doing something very, very interesting, very important. again, for an industry that it's not necessarily sexy at all. and a lot of people don't want to even get to it, right? But you know, that's that's that's the opportunity, right? and people refer to them as like boring businesses and

boring industries, right? But as you know, right, Elon Musk built an entire thing around a boring company, right? and it's it's one of the biggest movers, right, of products, capital and everything else in the in the world, you know, in manufacturing. so I applaud what you guys are doing. amazing founder. You have a great team.

as well and a great story to tell. So I'm I'm happy that you were able to share it today. But before we go, Carlos, I wanted to, you know, one last question. a lot of our audience, you know, of course is Latino founders, operators, investors, right? But specifically founders and operators, and they're building in different spaces, right? with the network founders usually have for the most part right

or other founders usually have. So when you f where you from where you sit right now, right, what would you tell the founder building something real in a space nobody's paying attention to yet?

Yeah, so I think mainly two things. So the first one is you need to really understand the problem and your market. you need to really understand what your client wants. So at least for me, I spent five years in the trenches doing tons of consulting for these companies. So it was like really understanding the problem, how what's the dynamic of the industry that I'm serving? Why are they having the problem that they're having? And what are

other solutions that they're going to before you actually build something. So I think that's the first one. And then secondly, as a Latino founder, I think it's really important to build openly in the sense of get out, build a network, leverage on people that already done this in the past and try to build that network that might be helpful for you to leverage on so that you can actually push your business forward. I mean, I know there are some great groups out there

That are just focusing on Latino founders. so I think that's like the other thing. build openly. I've done my startups like that. just this is what I'm building, trying to get feedback and then trying to build a network that can actually push me to the next level. I think network is everything. when you start meeting people like I mean, I met you, Angel and now we were sitting talking about this. And there's I'm always trying to help not only Latino founders, but

Definitely.

Carlos (27:03.03) as a community itself because you never know who are you gonna meet. so helping other peoples and try to get help that's always really useful f for any business itself.

Amazing Carlos. That's that's a winner's mindset. So I appreciate that. And where can find where can people find Corello just as a last thing, you know, website, et cetera. How can people contact you and get, you know, in a demo, for example, and things like that.

Definitely. No, that and thank you for this. they can find us in our website. So it's Corello.ai, LinkedIn Corello, and then my personal well, my work email it's Carlos dot Miguel at Corello.ai. Happy to answer any I'm I'm really try I get I constantly answer my email. So feel free to shoot me an email or a DM on LinkedIn. Happy to connect with everyone that has some questions or feedback again.

Amazing. Thank you so much again, Carlos. And for everyone out there listening to this, this was the corridor. My name is Angel Leon, the host. And today I had Carlos Miguel as he's the founder and CEO of Corello, which is a AI native manufacturing intelligence system for small and mid-sized manufacturers here in the United States. he's ba building this out of Boston, coming from Latin America, Colombia.

amazing founder with an amazing team building an amazing company in the manufacturing space. So I appreciate it, Carlos Thank you for being here today.

Thanks for having me, Angel.