100 Million Parcels Later: The Truth About Physical AI In Sortation

Tell the audience what does CubiQ do exactly?

If we have a like expensive equipment in the warehouse, we can like extract that much information in a single touch point.

How is AI really like changing the game? Here

we have one of the biggest databases of pictures, 3D pictures and metadata of every element that has gone through our system.

You did Shark Tank in Colombia. in Colombia and you won Shark Tank in Colombia.

What was the uh the experience like and was that valuable for for the business?

First, we were the company that have raised more money in this pro in this program in the history of the program until that moment.

Okay.

So, we raised a a lot of money with the highest valuation in the market at the point. Welcome to the supply chain technology podcast. I'm your host Navil and I'm so happy to welcome Camilo Garcia, the founder and CEO of CubiQ Robotics. Camilo, how are you doing?

I'm really good. Really good. Thanks for the invitation. I'm really happy for being sharing this space with you. Uh it was a surprise for me. I think this is one of my my very first podcast in English and yeah I'm pretty happy to be sharing this space with you. Thank you for for an invitation.

Amazing. Amazing. Yes. So Camilo, you'll tell us your story uh coming from uh from Colombia and and building a technology in Colombia

that is now uh very successful and that is taking off into the US. So uh before we jump into uh into that Camilo uh tell us about you you know uh your story and uh and uh we are shooting here from Miami but uh you know having developed the technology in Medellín in Colombia and now you know having attacked the US market for a couple of years. Tell us about about your journey.

My journey started like eight nine years ago. H the it was like a a hobby for me. I started like as a hobby because one of my friends invite me to his warehouse once and what happens there is that he was measuring everything using a tape and then after measuring the tape they were placing all the information in a in a sheet and after that it goes to Excel after going to Excel goes to invoicing and all that and was very prompt to errors all this process so it's like I think that there's a huge problem over here and I solved it like nine 10 years ago as a hobby. Then I start like doing a research in the market. Nothing was like great at the time, just one competitor at the time and I say like okay here here is a huge opportunity for for doing a business. So everything started like that like organic soap simple and then I also had another and like another companies that I was working on and then like I don't know six seven years ago I decide to quit everything I started working with my partner Felipe we focus 100% on the company and then start scaling up this this startup and from there like everything start running really fast we started with Colombia and then with another countries, a couple countries and then yeah as you mentioned we open market in United States as a like the main market for us and right now like 80 85% of our sales are over here especially in Florida.

That's great.

And yeah we are growing really fast. We have a a lot of happy customers. A lot of people that we are changing the way how they do logistics in their warehouses.

Yeah. And uh so Camilo has one of probably of the craziest career pivot that I know. So you know if if you go anytime to Medellín and uh and you uh you're looking for places to uh to go out. I remember he took me to uh to clubs and so on. So like can we go back to that? Like I would love you to share with the audience before we go into CubiQ. Yeah.

But like you were in the entertainment business like you had like you know like different type of businesses in nightlife and so on. How do you go from that to to a a technology founder and CEO in supply chain?

That's funny. Yeah, it's it's a story because I am I love music. I love music a lot. I'm a DJ too. I I have been

I remember you played on the you played when we were at that club in in in Medellín.

Yeah, I have been DJing for like 20 years already. And that passion for music took me to creating one of the most entertained like business in Colombia. Like we have multiple clubs, multiple things. And from there um everything started like that doing parties, doing some bringing DJs from outside, a lot of DJs and yeah, but at some point you get like really tired because everything has happened in the night then you have to work in the morning. So you have to quit one one of two jobs or you are going to focus 100% on nightlife or you're going to have focus 100% of being entrepreneur but you I think that those two activities are not like the kind of activities that happen very well together.

So I decide to quit everything on the nightlife. I have been like I still DJ at this point like once every two months something like that but not as frequent as before because I used to DJ I don't know like Thursday, Friday, Saturday, some Sundays. I was exhausting like I don't know a lot of party but yeah I really enjoyed I really enjoyed that point that point of my life but right now I'm completely focused on CubiQ. I love what I'm doing but also I need music in my life. I need a lot of music for for like balance for life balance. I need and also it's funny because I today is arriving one of of of synthesizer I just bought because I'm started to like doing music again my studio.

Okay.

So I think that I need that

in Colombia but yeah right here to to the apartment in Miami.

Okay.

So I think that I'm not going I'm not going to quit 100% of the music just to as as a promoter. I think I I already quit. But I think that music is going to be I going to be surrounded by music all my life. So

that's part and and you you have a good time over there. I remember that.

Yeah. Yeah. I had a great time. So you might think that Camilo is also an actor that he can do Pablo Escobar. I just told him like earlier before the the the podcast but uh but yeah it has a different different uh different lives different lives. But let's focus uh the podcast is on supply chain technology. So let's let's let's talk about that live and uh and maybe we'll keep some good stories for after. Um tell the audience what does CubiQ do exactly, right? Like for especially for nontechnical people or for people that don't

um

don't know really deeply the process and the workflows in the warehouse. explain to them how do you use physical AI robotics automation to actually provide the solution that you provide today and what do the solution actually do yeah for for these clients

I think that for a start like how is the process being done at the at the moment like how are classic warehouses doing this process and then from there like how we are changing that way so usually when a pro a box arrives into the warehouse usually has like three to seven different touch points. So each touch point could be like taking the dimensions, taking the weight, reading barcodes, extracting information from the label like hey who is the customer or like patch number, expiration date and all that they have to put everything manual from there. Also they have to take a decision that where is this package going is if it's going to a rack or is going to another warehouse you know like the sorting that usually is done completely manual um from there also they have usually to take pictures you know and extract the information that is with the paperwork and everything that it has. So as it can have like up to seven touch points, eight eight nine touch points and usually there's a lot of errors as in this process because depends of a person doing this manually and what happens there is that when you have errors it can increase the cost and the customer is going not going to be always happy with this. So we from there we created the first version of CubiQ what we did is a dimensional a classic dimensioner but I saw like hey you have a expensive equipment in your warehouse you can do a lot of things with that.

So dimensional like basically you come in you put a box or you put a pallet and the and the machine will take the dimensions.

Yeah basically height.

Yeah.

Height width length basically. But from there like hey if we have a like expensive equipment in the warehouse we can like extract that much information in a single touch point. So from there we create different versions but all the versions are like modular. So from there we are are now are able to extract everything that is on the label or everything that is readable in the box like batch number, expiration dates, customer names and all that. Also we have waiting systems for that. H we have actually recently we add added a new feature that we have thermal cameras for example for peri perishable uh merchandise that we can extract the temperature of of the boxes. Also we have h scanners that you can in the single in the same point like scan all the documents that comes with the cargo and from there also extract around 220 different entities of the of the pack of the of the paperwork like

so you cross the paperwork with the physical cargo to make sure that they match and so on.

Yeah. and extract for example what is the port of origin or what is the bill of landing all the information that comes with a bill of lading that usually digitizing that takes a lot of time also extract for example here in Florida is really common the couriers have like a caserero you know what it is like like I don't know how say to say that in English but it's like yeah like a cer that you receive the packages of different people in their warehouse and you have to know like this customer has a a like a PO box, you know, like a PO box that that of the end

you have to receive

unknown package is something.

So you have to read the label and from the label struct which is the name or what is the PO box number and do the match with the customer and from there like send it to Colombia or to different countries in Latin America. So we create a technology to strike that really accurate. Um the point here is that usually doing this process manually takes around 1 minute and a half in the best of the cases and also you have a lot of errors over there. Right now we are doing in less than a second capturing all this information with 99.5% accuracy. So that's what is CubiQ and after that we also create exorters and different technology to automatically uh basing on different rules uh send the package to different stations to different processes. So that's what we are doing right now pretty fast pretty accurate and right now we are a company that are is handling around 500 5 million six million packages per day. So it's a it's a a cons a really big number for a startup that two years ago was processing around 6,000 600,000 packages per day. So the growth of the packages that we are handling right now is reducing the the laboring warehouses a lot. So yeah.

Yeah. Yeah. And one of the thing that I think you showed me that I found was was quite interesting is that you you are a technology company that focus on the software but you also have a component of hardware. Talk to us about like

um you know I think you showed me some sortation capabilities as well. So you know explain to us how do you deal with these different things right like you're building the software that does the OCR the data capture maybe some of the of the um data transfer and all the intelligence around around capturing these different data points then you have the uh

you have your own hardware and then you also use some hardware that allow you to do sortations explain to us a bit more into into how how you build this together and why you've had to do it that way. Oh, that's an interesting question because the problem here is that you are grabbing a lot of information and the information alone yeah it's meaningful for companies but you have to take decisions with that information. So we decided at some point to start like building the hardware that requires to capture everything and also to take the decisions at some point. And what I started like more like in a way to adapt the product to the customer needs because at the end the customer has a different type of of of requirements but everything converge in the way that what do you do with the with the information that you have and every all the customer at the end says like hey we need to take decisions based on that information but not only sending an email or not only is like turning on a light we need to take decisions in real time. So that was one of the most u difficult uh problems that we had because it's like how all the data that you are capturing how you take decisions in less than less than two seconds or three seconds but you have everything on the cloud and you have everything in different parts of of the of the architecture layer. So how you send all this information to one side to another one in a pretty efficient way and also we are leading dealing here with a lot of pictures. So that requires a lot of bandwidth in the warehouse a lot of bandwidth in our servers. So it's how you keep everything flowing in a pretty smooth way in order to take the decisions in really short time actually to milliseconds. So that was most of that the hardware is how you synchronize everything in a in a way that there is no problem with the packages that are arriving to the sortation solution.

Yeah.

Yeah. And um tell us a bit about like so dimensioning there's obviously an element of shipping the right package with the right information or validating what you're receiving. But why is this more important? Is e-commerce is compliance some of the of the tailwind here? like what is driving adoption? You talked about five to six millions of packages, you know, like really a huge growth compared to when when you started.

So what what is driving that?

Why are customers like interested in buying this type of technologies?

When you go to any manager in the in this in this industry, there is one question that I think that nobody is able to answer correctly. It's like what is the cost of the air that you are moving or the cost of the air that you are storing?

Yeah. And I think that nobody is able to answer this until they have a dimensional or they have a digitalization solution in the warehouse. So everything starts like that because of the end logistics is a an efficiency problem is how you chip more. How do you optimize every layer of your of your supply chain as much as possible. So when you start like seeing that the dimensions are really important in all the layers like for example for invoicing if you don't know if you don't know correctly what is the volume that you are moving how do you know what if you what you are charging to your customer is correct or not.

Yeah. Or the same with the weight. For example, if you are moving like really heavy load, you can get a lot of penalties for having a load that limits that that are over the limits of a highway or something like that or for planes. For a planes, this is super important. And

yeah, because you get charge on chargeable weight and so and so weight plus dimensions are are critical.

Exactly. And for example, if you if you don't like send the correct data to your courier and they do dimensions after you send the ch the the ch the merchandise, you can also have penalties after you send the merchandise from your warehouse. So if you see like all the layers that are applying to the supply chain are really heavily affected by the dimensions and the weight. So that's why you recently see a lot of adoption of dimensioners in the industry and also because if you go and you compare like the time that a human takes to take like a approximate dimension of any cargo is like 10 20 times more of what a machine takes. And also what happens when you go to the data is like over the time uh an operator that is taking manual dimensions of the cargo goes about 60% accuracy versus the machine over two months. The first five days they are super accurate. They put a lot of attention and everything but after a week two weeks that you start seeing uh how the accuracy data starts going down and that's affect a lot any business. Mhm.

So for example for airlines is really important for couriers freight forwarders are really important this data for like e-commerce is super important also for example going to for the base of the of the e-commerce industry if you don't know what is the the size of the cargo that you are going to pack after you are doing the the pick and pack of of which is the most accurate or the most efficient box that you have to pack this uh I order that you receive, you are not going to be efficient in in the way that you ship your cargo. So everything you see any all the layers that comes with the supply chain is heavily affected by this.

Yeah. No, and it's interesting because it does apply to air freight does apply to sea freight especially if you're doing like less than container load. It applies to LTL. So it applies really

every every Yeah. every layer, every like supply chain is really affected by that and and when you add all the technology that we have a part of just dimensions, you are getting a lot of information from the cargo that you're moving on.

Yeah. And I guess with machine vision so so it's perfect segue. So talk to us about this, right? like talk to us about so um now you are grading every time more data and and the camera systems and the the software layers or the the AI layers that you're putting on top are every time more sophisticated. How has this changed what you can do actually with the pictures with the documentation that you're processing you know structured versus unstructured data how is AI really like changing the game here

in our case I see like there are two things that we have been working really hard in the last seven years. first is that we have one of the biggest databases of pictures, 3D pictures and metadata of every element that has go through our system. So right now we have a database of around 100 million packages in our database. So from there we every time that we want to train a different model for identifying different things in the market we can do it pretty easy because we have a people in the in our company that are just focus on training and tagging the images. So from there we can train for for example recently we we came with um solution to identify packages that has damage but any kind of damage into CubiQ line. So we grab I don't know like four or five million images of of packages that comes with this condition and right now we are able to identify all the packages that comes with any damage into our system in less than one second. So that's really good for the industry. Also we're able to identify batteries that like all the stickers that comes with batteries and all that we fragile or any type of

yeah all that kind of cargo. Also, we are now right now we just releasing a new model that usually with 3D cameras there is a problem that some kind of materials are really hard to to sense in a proper way especially like this camera because we use different type of of hardware in our technology. So we are using LiDARs, we're using time of flight, we are using different type of 3D cameras that the combination of this help us to be super accurate with the with our technology. But recently we have seen like combining all these and also using the color images for reconstruction is helping us a lot to do a complete reconstruction in some cases that just one camera are not able to see like for example backstretch. So we we used to you we used to have like to solve this using super expensive hardware LiDARs. Right now we are reconstructing this with with using the color image that comes with the with the image and the the result that we are having right now is really good. Also the AI that we are using for for the documents we are using AI to identify the documents and extract extract the entities of of those. For example, if you put a paper or like a bill of lading invoice or packing list, we're able like to identify all the entities that comes with with um with a document. So for us like we are using that for also for that and recently we are trying to to make um to solve a new problem that we have seen in in the sorting solution and what is this is like usually sorters have been work in the industry forever with rules. You have rules that you have to put rules and based on this the sorting works. Now we are changing that to have context instead of rules.

So we are putting context into the package in bas based on the context using modern AI to do the sortation rules or not the rules based on the context. So you just put like uh like an LLM on on the middle that helps to do the sortation of the of the packages. So we are changing right now using AI for having context on the sorting solution in a way that I think that is going to work better because what we have seen is like what happens is that when a package gets out of the standard rule that you are expecting to work everything fails. So it's how you make the systems like to to be fail safe or not fail safe but at least to have more to increase the rate of successful sorting that's where we are working right now. So give me an example for example where would context where would AI LLM context um would that be in the case of like um a sortation process where for example a package come in it's going to the right destination you will sort it into one of the the destination or route but it's damaged and consequently

yes

that will be put aside because it needs to be reviewed like

but usually when it's damaged what happens there is that it's not going to read a barcode right So it's like hey you don't have a barcode information of the package so usually this is going to fail fail fail route but it's how you using AI you expect that to change to a positive result. So you can do a reconstruction of the barcode you you can do like a lot of post-processing of the images in a way that instead of being as a fail goes into the right position.

Understand? Because usually a sorting solution, yeah, you have a lot of packages going through, but there are a lot of packages that comes with a wrong barcode, wrong label, damaged label and these situations that makes like six, seven% of the cargo that goes to the failed station. So it's how you use AI in a way that you can improve this. So that's what we are working on.

Yeah. Yeah. Yeah. No, and I think the example you gave on u matching the data from maybe from bill of lading into um into a damage back code. I would I would imagine that that's an example where you actually don't need to put the package aside and say this is a non-readable package, but you could actually

uh match back the data. That would be an example. No, that's that's that's really interesting because then you can really improve the speed and the operations as opposed to just just capturing more data across across the uh so um tell us a bit about the the the journey also. what um what are the type of customers that you focus on a that are today clients of CubiQ a um you know a 3PL um like like what's what's the the customer usually profile that buys this type of technologies

so we have like five four verticals that we are like completely focused on one is airlines airlines is one of the the verticals that we are attacking right now the second one is freight forwards freight forwarders all over the globe. We right now we CubiQ is in a company that has operations in 10 countries. So we are from United States to Argentina and we are open in Europe right now and from there is like freight forwarders airlines 3PL's couriers uh we had also CPG we created custom solutions from CPG and I think that's that's main all the industries that we are covering also e-commerce but e-commerce you have the or the companies that are doing e-commerce by themselves but yeah that that's those are the industries that we are covering right But the point here is that it's really funny because all this industry they're moving cargo right but at the end each company has a unique problem that I think that's that's when technology like the experience like plays a really specific role in solving this solution because we have like I don't know we have like 10 different airlines but the way how they use the technology is completely different in one of each one and is how we solve that and how we use our creativity and also we use our team to solve all these problems that they have and also the same applies for freight forwarders for each freight forward has the same problem or usually you think that they have the same problem but it's how you solve that in different way so that's that's that's funny

yeah no like give us like give us an example of this like how do like um so like give us an example on the airline like how would an airline use a techn technology versus another one. Is it like one does it for parcel, the other one does it for like cargo or like there's even more to that?

For it's more that the way how they use the information. For example, one is using it for for charging the right price for to their customers and other ones are using it like for having like optimize the way how they load the plane. So they don't care about charging the correct dimensions to the customer. They don't say that's pennies for me is how do you optimize your plane and how do you optimize the weight inside your plane

because of the end like if you have the right the wrong the wrong um

balancing balancing of the plane it's going to affect a lot. So they're using our technology for that for example. So for me, yeah. And how do you send the information? Because for example, this is it's pretty funny is how do you calculate the center of mass of the cargo that you are receiving based on weight and the contexture of the cargo? Because at the end we are a dimensioner that not only loads or not only gives the length with the height, we give a 3D context of the object that you are dimensioning. So it's how you based on that information you are able to to provide the center of mass of the cargo they are receiving. So it also helps a lot to to airlines for example in this case.

Yeah absolutely. So Camilo you touched on AI and some of the examples and how it's impacting the business and and your techn uh huge hype around AI in supply chain. Do you think that this is actually really going to transform physical AI and the areas of business you're tackling and the wider industry? What's what's your take on this?

That's a that's a good question because for AI to work first you need a lot of data. So what is happening here is like for LMS the data was available in everywhere. So you just like scrape a web page and you have all the web page information. You scrape Reddit and you have a lot of the information. You scrape GitHub and you have the code. But in the case of logistics, most of the companies of this information is completely private. So the way how is like learning these algorithms depends more on the way like how companies are like capturing this information. So from there I think that there are two things that are happening from my perspective. One is how the process of the logistics are being done. So in case that for example the way that we come to the to the beginning is like logistics works pretty well until something goes wrong in the chain. Something goes wrong and everything is like a complete chaos in a logistics company. know from the shell like you create systems that

are created like fail safe but in case that a package is going to Poland in some and in some way appears in Germany there's a lot of things that have needs to happen to that cargo to cos back to to to Poland and usually it's human related is a human that has to relocate the package do all the customs manually and all that things that have that must happen for My perspective like what is going to happen with all this a boom is that all this process that usually goes into cows you can create agents to solve part of that

for customs for h for a lot of processes that from my perspective can be automated from from this but from the physical layer there is a lot of technology that's coming every day but I think that we are like I don't know five seven six years a way to have a solution that at least can solve most of the problems because you see like the robots that are going out right now are focused on a specific task.

So there is a road that is good for for um roing or for loading the conveyor belt, the one that was released like one or two weeks ago that the it was able to achieve the same speed of a human. But

you're talking about the video figure that was doing the pel induction in the

Yeah. Yeah. But if you see like it's really specific packages.

Yeah.

When you go out of

it's a very simple process too, right? It's a very it's a very simple process.

But but you know don't see right now like if you say that robot to move a a pallet is going to fail. Absolutely. So are they created and they are trained in really specific scenarios and that's for example in our case the thing that took us more time because you can create a dimensioner. You can create dimensional in one week but it's how you create a dimensioner that is able to process boxes but also is able to process box how is able to process any regular shapes how is able to process in different light conditions and different warehouse conditions and that's what is difficult in this industry because you have a lot of different scenarios that is completely difficult to train to graph data and that's where I think that Most of the companies are focusing on you can creating solutions for a specific scenarios and that's what's going to happen in the next years. You are going to see the robot that is specific for for palletizing for grabbing in boxes and all that. But in a robot that is global process about label is going to take a lot of time for my perspective I think. So uh but for the process that there are a lot of hardware software working in the background. I think that what is going to happen is that what I say agents that are going to solve a specific problems when the package or the cargo goes away or moves away from the process that was created for Yeah. Yeah. So you you you really see this as like a I think you said something that um is really important is

the data and the context of this rules of business is not there right like so you have to build that right you have to build that knowledge to be able to train the agent it's not like

you're going to do like a query on any type of LLM that is a public LLM where you're going to get the business requirement of

American Airlines cargo or or Delta Cargo or any of these companies nor any free PL publish that information. So you have to build that knowledge and that's why I think you were talking about your data and the amount of packages information you have as definitely a leverage towards building the agents towards that. Is that is that the right way of of understanding?

Yeah. Yeah. Yeah.

Okay. Very good. Let's let's shift a bit gear on uh you know on the technology and you have started the company in Colombia in Bin and then you've decided to expand you expanding to other markets in Latin America and then to the US. Um tell us tell us a bit about the journey of a founder that um you know is not you know born and raised or have not like like evolved in the US opening a company here pitching American-based businesses and so on. How how is that a and how do you position technology in an industry that is also uh has been harder to penetrate than than uh than maybe other industries? That was incredibly difficult. I think that was most difficult to get into US market than creating a technology itself because in the industry usually you see like the standard are German companies, American companies or Japan companies in robotics that's the classic that's the usual

and you see

more and more Chinese too.

Yeah. I'm more Chinese recently, but when I started, the Chinese was were just getting started. But in my case like hey comes a Colombian guy with 20 something years to a warehouse offering a solution that is going to solve the problems that they currently have and they s like you know like the first problem that I have is credibility and that's pretty basic like how a Colombian guy is going to solve this problem and also I have a

you know that you know that still today when I say that you know I talk about CubiQ and I say yeah this company is based out of Medellín but they they operate in the Still many people including Colombian people are like really

it's still happening but this but right now I have a ro a roster of a lot of companies that are believing in me and a lot of successful stories that helps to

to change that pretty easy but when I started I have I don't know two companies in logos in my in my web page and they say like okay how I going to pay this huge amount of money for a hardware that I'm not sure is going to work whatever so that was the main point. The second I have a joke that is that is yeah like a local joke because first Colombia is not the the the

it's not the robotic capital of the world.

It's not robotic and and itself is not the entrepreneur capital of the world. Creating a company in Colombia is incredibly difficult. You have to go into a lot of step steps, a lot of government issues, a lot of bureaucracy that you have to go through. And for me it's like passing Super Mario in super difficult level.

That's the point of creating a company in Colombia. And anyway, we are here today like expanding to United States. We have a lot of customers in United States. Um yeah, that's the point here is like you believe really hard really hard in yourself and you think that you can kill it, you can do it. And right now with all the all the like helps that exist like LLMs with open AI with all these technology that helps to create technology faster. I think that the gap of creating technology in third world work countries is going to be reduced. But when I started was like really difficult but for me the credibility was the point of inflection. How you change the credibility? But there were a lot of people including you that believe in me at the point and we are here right now changing in some way how logistics is made in the world

was credibility the biggest challenge more than technology you said right

I think so yeah absolutely

for the US market yeah

it's two things credibility and capital

for example for hardware for hardware the capital is pretty this is a business that is really capital intensive

so in our case that was one of the biggest limitations scalability and capital.

Yeah. So that's another thing, right? Like around fundraising and around companies. So you have the uh you have the ability to just go software only and some companies do that, right? But when you go and you start to touch hardware also there are some investors that get scared of that. They're like yeah we don't really obviously now also in the age of AI having hardware give you some level also of additional defense defensibility. Yeah. But what's what's your view around you know should you if you had to go back would you do it differently or you would be still still thinking like no I need to build the full mode and business and solution here I need to have both I need to control hardware and software to make sure that the actual solution is is is uh valuable and it's delivering the outcome

of the end what happens with software is that you can scale easily you have a web page you have an application you just can deploy anywhere in the world with just one remote connection with somebody. But when you add the layer of hardware, it's incredibly difficult to escalate because what makes the the business difficult is like the amount of layers that you start adding into the for example the supply chain, the type of sensor, the technical support that you have to that you must have in any country that you're operating also the regulations that each country has. There are different regulations that you have to comply when you are using hardware. So those layers makes the business not so attractive for VCs because if you you go to the VCs right now like 95% 95 96% of the VCs that I have touched they say they don't invest in hardware because they prefer the classic software business but if you do the same right now software has been commodized so you can create software in a week using just by coding and all that pretty So this the hardware right now is like the premium on the on the industry because creating hardware first you cannot commoditize that because you need to create the hardware itself. You need to like for example in our case when we are we were creating the rollers for the conveyors. We try different amount of combinations of the rollers and how we increase the life the lifespan of the rollers and that requires like that require from our side like 40 50 different um h changes on on the interventions yeah interventions and innovation in the roller itself. So it's how do you get into the the liability of the hardware requires a lot of effort and you you don't get that in in a question in the in an LLM because if you change one degree in the way how do you make me manufacture a roller is going to affect completely the result. So I think that right now the the hardware is the one that is going to dominate the world because everything is heading to that way. How do you automate? How do you remove humans? How do you make any process any more efficient and more more accurate in the way? So I think that for me for example to business was really difficult. A Colombian guy in Colombia the headquarters were in Colombia United States is fell in United States you know like I had all the all the you know is like it was

you had all the red flags.

Yeah. all the red flags for

and on top you had like the Pablo Escobar must touch.

You had the Pablo Escobar vibe. So for for VCs that probably thinking like maybe we should launch another type of business or maybe we'll invest in another type of business.

No, but yeah, but it's I think it's great because I I think it's also great to share your experience because a lot of founders there and we really intended the podcast to be uh international, not only about US entrepreneurs. And that's why I'm so happy that you're the first international um guest. Um

I feel sometimes there's a lot of international founders, LATAM, but also um in Europe or Eastern Europe or you know places in Asia that they look at the US market and they're like well

you know somebody already done it there. It's going to be really hard you know. Um and it's more of a mindset thing rather than they actually have low lower level of sophistication on technology or that they are less capable than people here and you took the bet right and so if if you had to do it again like same again the same question on the US market what what's your take on this and what would be your recommendation to other entrepreneurs that are international entrepreneurs around the US market is it worth it yes or no and would you do it again in the same way or what would you would you do friend.

I think that you're going to focus on estate, you have two options. you can pour a lot of money into like growing that but that depends of two things like let's let's talk about hardware companies okay because it's software you just put a lot of ads and you start like taking customers and I think that's that's pretty easy but when you have hardware like any decision it has a cost associated for example in our case what we did we focus on a niche in a in a complete niche and from there our niche was freight forwarder in Florida and we focus 100% on that at the beginning. So right now I think that we are the leaders here in Florida for for dimension and for automation like two of every three warehouses are using our technology that has automation. So from from

and here there's a lot right I mean like what is like 3,000 or something like this. So if you go to any warehouse and you see automation that most of the technology that you're going to see is CubiQ but from there is like this is a really competitive market absolutely competitive market and one thing that we decide is that we are not going to compete with price. we're going to compete with the best product and with the best technical support available because technology can fail but it's how you solve the the problem in the most efficient and the most professional way. So right now we create a technology that I think that's what most of the customers that use CubiQ love is that in case that CubiQ fails, that may fail; usually the problem is solved in the next 10 minutes and that for robotics is uh it's you you know it's like uh a goal that is really really difficult to achieve. So we focus 100% on that. How we provide the best technical support available in the market. How we synchronize all the all the people that we have in the company for that. And what's happening there is that we have when you see for example an airline in an airline the system fails. There is a plane that cannot uh take off on time. If you see a warehouse, if you fail the technology, it's going to be a truck that is going to be parked a lot of hours in the warehouse or it's going to leave with a lot of barcode without reading or it's going to affect a lot. So, we focus first instead of the technology, we focus on how to solve the problems and then we start building this technology that is on the back and that's I think that what the US was that's what the people loves more of CubiQ and that's how we decide to grow. So and definitely I will focus 100% in United States again is a market that is really beautiful because for example in in Latin America we we talk about transformation how you transform the the process in United States the conversation is completely different because they are very transformed they are already super efficient it's like what do you going to offer in a way that they took their technology so the conversation is completely different in in Latin America transformation is is increase the efficiency, increase the the warehousing capabilities. How do you reduce the hand labor? Hand labor here is incredibly expensive for example.

So we focus a lot of that because using CubiQ for example you can reduce the hand labor warehouse around 80% in the process that we are affecting. So, so yeah, it's like but it's what you say it's the mindset. You have to have the mindset of competing with the best companies because everybody wants to come to United States. You have to

Yeah. I mean the addressable market is bigger. I think the willingness to take risks is is higher

and the um the importance I mean you didn't maybe touch on that to that same level but I think there is something there also on what about speed right like because obviously when you do you know and you talked about this at the beginning around the manual processing and people coming literally with a measurement manual measurements and paper and so on obviously that create bottlenecks also and the speed of processing completely different so how how important this speed for the companies you serve around not only package processing but also data processing a a value proposition

but at the end everything translates translates in ROI everything everybody here focus on ROI yeah

what is your investment how fast are you going to return that investment and everything focus on that so that's why

show me the money

yeah show me the money and how you are going to recover that as fast as possible so instead of speed the conversation talks about how you going to recover the the the money. In our case, you have like multiple points that you're going to recover the money because you are going to have the right charging weight to the customers. You are going to reduce the hand labor. You are going to automate a lot of process that you have. You are going to reduce errors. You are going to have everything digitizing your mess very fast. So in our case was pretty easy because usually the arrive of our products is the longest that it took is is like six months seven months something like that. So when you go to the right person and that take the decision is pretty straightforward is say why I'm not going to use that and that's why when we start changing our business model to robot as a service because we saw that reduce the friction with the customers and at the end they are recovering a lot of money and having the highest ROI based on the investment. So that's what we do.

Yeah. No, absolutely. Give us a sense like so you talked you touched on the money thing like give us a sense what's what's the what's the type of ROI that people would see with CubiQ a even if you count it from a pure human labor perspective to

CubiQ as an automated system, or any type of way that today you calculate ROI when you engage with these customers give us a sense what type of i people can expect in the US market where labor is more expensive maybe and if you want to give a comparable in another market

that depends on the solution But usually we have solutions that goes between $5,000 to like $1 million depending of like everything that they also depends of multiple things depends of how many people they're using in that process how many shifts do they have because there are companies that are working 24/7. So you obviously when you have 24/7 shifts the ROI is going to be higher. Yeah. Also if the weight chargeable weight is implicit in the operation because there are companies that they don't have chargeable weight and usually this is one of the points that affects more the of one of our our solutions and from there that based on those three combinations it goes between three months we have Philip three months to six months the the classic of our solution so we have

it's very quick

it's pretty quick to six months for example in Colombia we have one customer one career company that actually they did we did a ROI calculation of eight months and when they start rolling 100% with the solution it took like four months of of so we reduce at half the expectations of ROI for for the solution so and for us it's like it's pretty beautiful because you expect some things to happen but and also we have like a conversation we we actually try to underachieve when we when we say like something we try to don't put like super optimistic numbers in in the in the presentations we try to actually to underachieve the what we are expecting and we usually what happens is that we

to underestimate to yeah underestimate everything and at the end we provide like better numbers for the customer and that's that's

and that's always a better way to do it right to go with a conservative numbers and surprise them and I mean still like still even in 8 months in most of the if you compare this to most of the technologies that you have out there

it's you know maybe with the exception of agentic AI it's it's uh it's pretty fast right like I think in agentic AI you can get to much faster our I as well but but uh but in anything that is physical I think it's a very it's a very short timeline so so that's that's really good

you talked about the expansion into Florida as already like well so what what's next for CubiQ where are you going like what's the the plan of expansion either geographically are you launching new products new solutions uh or really expanding what you have already geographically

so we we created a new product last year I think that was most one of the most difficult product that we created that was the DTPS there was the the drive through the scanner that we were able who

so explain to the audience what is that? So

you have a you have a forklift that usually you have the cargo in front with the with the forks and what happens here is that we create a portal that you go through and we are able to weigh, take dimensions and read all the barcodes and all the information of the pallet on the side and the front.

And the point here is like how do you take the dimensions of a cargo that is first in movement second how do you read the barcodes also in movement. So you have to reduce the exposure of the cameras a lot to have a really good illumination of the area and how do you process that fast as possible. So we are using Nvidia hardware. We have been using Nvidia hardware for since day zero of the company for like nine 10 years and we are using GPU for the processing since day zero. So that helps a lot. That'll help us a lot to to create this kind of software that works pretty well in this in this type of of of environment. So that's what we created last year that we released it and we are expanding that product worldwide because the solutions that exist in the market doesn't like work like we think that it might work in in the conditions that we are creating right now. Okay? Because if you are not able to read all the barcodes of a pallet that is going through it makes no sense to have all the compute solution in my case but what I see also um

that just on that one like there is there is existing solutions with big towers and so on. Um

what's the differentiation there is that you can do it better or at a cheaper cost. What's what's the differentiation compared to what exists? There've been some solutions that

that uh at the dock

put two type of towers which are quite expensive but like what would be the difference?

The first is that they offer just the dimensioning. We are offering the complete solution. We are able to read OCR. We are able to read barcodes and we are able to capture everything.

We are able to count the amount of boxes that is in the pallet.

Okay.

We are able to extract batch number expiration dates. So from the pallet itself and obviously the cost is much lower because we have like using a different type of hardware that in this case reduce the deployment cost a lot. So that's I think that's the main difference that what we have seen in the market right now.

Okay.

Yeah.

Okay.

From the from the company perspective from the market I think that we are going to expand anymore. We are in 10 countries right now. Maybe next year the plan is to complete expansion in in Brazil and have the com the the the entire continent covered already. But I think that we are focused on this in these countries no more to have the right technical support to have the right people in these countries but I don't think that we are going to spend any more at least for the next two or three years. Um yeah that's pretty much what we are focusing right now. Okay, great. Um, as we go into the a bit of the wrap-up, uh, you have a another funny story around the UD Shark Tank in Colombia.

Yeah, I did.

Shack tank in Colombia and you won Shark Tank in Colombia.

Tell us about that experience. What is it like to be on that stage uh, pitching your company? Was that helpful? added some credibility just a good marketing like tell us like what was the uh the experience like and was that valuable for for the business?

It was funny because yeah from different perspective I I think that is really good for a company from the PR perspective.

Was it fun? Was it fun?

It was a lot of fun. Absolutely. It was a lot of fun because that happens right after the pandemic. So So everything was like starting to move in again. I have to do the the pitch with with a mask and then you have to take it out just for like when they start recording after they cut you have to put the mask again. It was funny that um and after that what we we did with our marketing team is that first we were the company that have raised more money in this pro in this program in the history of the program until that moment.

Okay.

So we raised a lot of money with the highest valuation in the market at that point. And second one we create an entire strategy for for when the the the chapter go went online we did a a merchandising strategy for that. So all the media was around us at the time. So for the credibility in Colombia for example that gave us a lot of leads and we closed a lot of deals on that moment and also yeah that that that was was really good for CubiQ I really enjoyed enjoyed a lot. Yeah

that's great.

I I don't like a lot a lot of cameras but at that moment was I think that CubiQ need that needs that credibility in the market so I decide to go to to the program for that.

Yeah. Yeah. No, and I really like what you say about credibility because when you touch especially when you touch physical automation.

Yeah.

Credibility is extremely important, right? People will look at you as show me what you've done before. Who do you work with?

Yeah.

And so not that TV or Shark Tank would change that, but it does give to the people a sense of it's an established company. They have some real solution that are actually already in place and that are operating. And I think think for anybody out there especially if you build something in the physical space that have high risk also and you can lose a lot of time that point on credibility is really important.

Yeah absolutely.

Would you agree with that?

For robotics is everything I think that the success cases I think that is more important than everything. Uh because one one I think that in logistics everybody knows everybody like you you are one phone call away from any person in logistics because everybody knows. So when a company is providing uh bad product it expands over the people pretty fast and in the same way for su su success cases. Yeah,

in our case like two of one of the like three one of of one third of the of the leads that arrives to our to our web page or to our marketing team are people that are referrals from other people that are using our technology. So that's for me that's one of the biggest KPIs that I handle because that says a lot about the product like the product is really working the product is really giving the leverage that we are promising that is happening over there. So

yeah, of course. Yeah. No, congratulation on that. It's it's not that that uh it's uh not that easy in the sense that you're tackling a complex problem with a lot of dynamics and also in the physical space, right? So amazing. Camilo, where can the where can the people find you and where can they follow Cubic and uh and get in touch if they're interested?

First the web page cubiqnet.com. Cubic is our

we'll put it into the we'll put it also into the

and also in LinkedIn uh Instagram, Facebook. I think that we have all of them. Cubic cub iq

and you can follow me in LinkedIn as Camilo Garcia.

Yeah. Yeah. We'll put also the the the link for Yeah.

for LinkedIn. Well, Camilo, thank you so much for coming. Congratulations again to you, Felipe, and the entire team of Cubic, Alejandra, and and the full team um in the US, in Medellín, and across the region. We wish you a lot of success for the future and we're looking forward to have you again and to to hear about how the company to expand and develop. Yeah. Thank you so much for coming.

Thank you. No, thank you. Nav has one of the persons that has helped me a lot in into this journey.

It's like my a part of my mentor is like the person that I call when I'm like, hey, I don't know. I I I need to solve this. And you are always being there for me. And I really appreciate that. I really appreciate that for her always like being there for her everything that I have to say. So thank you Navil.

Yes absolutely not anytime. So yeah it's funny because we didn't know each other but we met a really like through the the ecosystem of the industry and so on and uh

and uh we we we have built an amazing friendship. So anytime Camilo you know anytime I can be helpful we'll be here. Yeah great to see you again. And all the best to the team and we'll put all the links. So if you want to connect with Camilo with his team or just follow them and see what they are doing. Dimensioning technology cubiqnet.com. Thank you so much Camilo for coming. We'll see you soon. Sure. Cheers.