Defense Tech Spotlight: Cyvl.ai
I had the privilege to speak with Daniel Pelaez, Co-Founder and CEO of Cyvl.ai. Here are the highlights of our conversation:
What is Cyvl.ai?
The core mission and vision of Cyvl.ai is to help the world build and maintain great transportation infrastructure. We are achieving this through digitizing physical transportation infrastructure with sensors and building enterprise AI solutions for governments and infrastructure owners/operators to use.
What is your background and how did you come up with this idea?
I have a technical education, I studied electrical engineering in school along with computer science and robotics. The first job I had was back in my hometown for the public works department, a classic summer job with a lot of manual labor. That was my first experience seeing how inefficient local governments are when it comes to managing their physical infrastructure assets. They have no software tools, and limited data and information, so it felt very disorganized and like they're barely staying afloat dealing with all the issues they face. I saw the ugly side of what it's like to run a small city and maintain all the physical infrastructure, which gave me an understanding of the root of the problem.
I realized that these guys have no way of regularly updating their data/infrastructure reports, and I went back to school with that in the back of my mind. In school I learned about LIDAR technology applications for self-driving cars, robotics, defense, 3D mapping and thought why can't we take this awesome technology and build a tool for these local governments to create and manage a data catalog of their physical infrastructure? The idea was to slap the hardware on a public works pickup, let them drive around the town to collect the data and we will develop processing algorithms so that every time they drive around, we are updating the digital map of their infrastructure. This data will help the government understand what it needs to fix and how to prioritize the roads that need to be repaired.
What data are these sensors capturing and what are you using this data to do?
When the sensor is driving around it is capturing as much raw sensory data as possible. It captures 360-degree high resolution imagery, like Google Street View, and 3D laser scans from LiDAR sensors which are synced up with the imagery as well as high quality GPS information. Once our customer captures the data, they upload it to our data processing platform and that's why we're really a data processing company. Then we are running the data through tons of algorithms to build a decision-making prediction based on the information the sensor has gathered. The algorithms and AI help us determine the difference between a sidewalk and asphalt road, it can tell us what roads are painted, what roads have signs, what types of signs, etc. It's doing an inventory and determining locational data as well as assigning dimensions to everything, so we have accurate data on how tall this sign is or how wide the road is. Then it will do prescriptive condition assessments, so we will know if the roads are in good or bad condition or and where all the cracks and potholes are. The final product is a report we can use to tell our customer this is the order you should pave your roads based on your budget and our conditions assessment.
You keep talking about the roads in particular, is that your primary focus when you refer to physical infrastructure?
The assessment of roadways and paved surfaces is the number one solution we provide to our customers, then we have a whole list of add-ons, for example they can also gather data on the signage or the trees or the sidewalks. But roads are really the common denominator for all infrastructure management across every town and city in America and abroad, everyone has them, everyone needs to use them, and everyone needs to maintain them. In the last year we have expanded to collect and report things like signage for cities. We’ve found that cities are very interested in pedestrian safety and mobility, so they want to know about their sidewalks, and ensure they have proper crosswalks that are visible at night. They also want to know about bike lanes and bus lanes so we gather that data and can build reports on it too.
You briefly touched on this earlier, but how big is this industry?
Starting with the market size, look at what is being spent on maintaining and inspecting transportation infrastructure, the roadways and highways, and the air infrastructure, runways and airports, and then rail. It’s something like $6-$8 billion worldwide just on the inspections and capturing data on these assets, that doesn’t even include the actual construction costs. In the US it’s around $2-$4 billion dollars. We see plenty of room to grow, there are massive inefficiencies that exist in this system, so this industry is ripe for disruption. What’s great though is that it’s not even disrupting, the people we are working with desperately need us. The civil engineering firms have labor shortages and a massive backlog of government jobs, they need someone to get the work done faster.
Are there competitors doing this or are you the first to disrupt the way infrastructure data is captured, digitizing a process that still uses pen and paper to gather data?
That's the really interesting thing, we're really in this wide-open space. It's not the sexiest, most glamorous industry, people are usually not thinking how can I disrupt the public works, roadway, infrastructure industry. That being said, there are several other early-stage startups trying to go after the same thing. We would argue we have a very differentiated approach with our sensor fusion between the imagery, LIDAR data, and some other data sets we capture as well. Some of our competitors are building a camera only solution or mount an iPhone to your car and we can capture it that way as well, but we have a high-quality data set at a very affordable rate, the best of both worlds. There are other companies that are strapping millions of dollars of sensors on vehicles to get very high-quality information, but that requires a huge hardware investment. We don't charge for the hardware, and we're going to give you incredible process results and reports on your infrastructure. So, we feel like we're at the top of the pack and leading this industry disruption. For now, we feel pretty excited about that and as long as we can continue to execute on our growth plan, we will become a household name.
What is your business model, how does Cyvl.ai make money with this technology?
We have two models, the first one is like an annual subscription model where our customer purchases a certain amount of data processing credits. They can gather data with the sensor as much as they want and over the course of the year, then they’ll use their data processing credits running that data through the software to develop the infrastructure reports. We have a second model tailored for companies just starting out with us because they won’t know how many credits they’ll need to purchase. We will charge a small rental fee for the sensor, and then they'll upload the data, and we'll charge them for each report we build. It's a pretty easy, simple business model that our customers understand and makes it easy to start working with us.
What's been the most challenging part of this whole journey?
The funny thing I realized, and I think a lot of entrepreneurs realize this too, we think it will get easier. It's a good joke, because things just keep getting harder. For example, when you hit a massive milestone, like a big sales target or you release an awesome product, that just sets the bar higher because now people are expecting that. But with all that being said, I think making that first sale, that first paid, signed contract that I think was $1,000 for a one-mile test pilot with an engineering firm. That $1,000 felt like the biggest checkmark of yes, this idea is going to work that I think we've ever felt. We spent an entire 6 months building a prototype and cold calling hundreds of people, towns, and cities trying to find someone to say, ‘OK, I'll give you guys a shot.’ Somehow that first contract opened up the floodgates and we were overwhelmed with business those first six months, which was crazy, but it was a really good problem to have.
Building the right team has also been challenging, and it’s very important to us. We've tried to be very, very intentional about who we're hiring because when you are in these early stages one bad apple or someone that just doesn't mesh well with the team is 10%-20% of your company. That’s why we have spent a lot of time making sure we're building the right team and finding the right co-founders.
Read our full conversation here:
