The Future Of Vehicle Inspections Is AI-Assisted With Eliron Ekstein
Let’s face it: car inspections suck. They take too much time and can be very subjective. If we want to make the process more efficient, then AI-assisted vehicle inspections are the way of the future. Chad Burmeister sits down for an interview with the CEO of Ravin.AI, Eliron Ekstein. Eliron talks about his inspiration for the creation of their car inspection app, and how he cultivated a passion for cars. Listen and learn more as we discuss AI use cases in unexpected places.
Listen to the podcast here:
The Future Of Vehicle Inspections Is AI-Assisted With Eliron Ekstein
I'm with Eliron Ekstein. He is the co-founder and CEO of Ravin.AI. There's not a lot of other .AI companies that I can say I've talked to so it's a privilege to have you on the show. Eliron, thanks for being here.
It’s my pleasure. Thanks for having me.
I love to understand before we go into the use of AI, I like to connect you with our audience because we're all people that are reading this conversation. Tell us about when you were younger. In your early childhood, where did you grow up? What was your passion in life when you were a kid?
As a child, and this is pretty much like predicting the future, I remember myself sitting on the terrace and looking at cars driving by. I clearly remember that guessing game. I was sitting there and looking at the cars driving and I would try to figure out their make and model. Back in the day, traffic wasn't as intense as it is now. It was maybe once an hour, but I remember myself eagerly waiting for the next car to drive by. Soon enough, I was probably fluent and I was figuring out most of the makes and models in my neighborhood. As we started moving homes, I spent a good chunk of my life traveling between places, the challenge intensified. Looking at cars now as part of the business is probably a self-fulfilling prophecy for me.
I was just with my son a few months ago and there was a car show. We pulled in, and there was the same model and probably year of a Cadillac that I drove to my prom. It was old then, so it was the 1960s, 1950 or something. I told my son, “Did you know your grandpa can actually tell you the year, make, and model of that car?” He's like, “You're kidding me, dad. There's no way.” There are so many types these days. I went over and he's like, “Prove it.” We went over, took the picture and texted it to dad. Within minutes he comes back and he tells us exactly the year, make and model.
I've never met anybody in my life who has that ability and now I have, so that's pretty cool. It's funny that sometimes our passions become what we enjoy doing later in life. I think the people who figure out how to pull the cord and pull the slack out and cause those two lines to connect live a very fulfilled life. I see the picture of the car behind you. Tell us a little bit about Ravin.AI. Looking at your website, it seems like a really amazing technology that you've developed.
Sometimes our passions become what we enjoy doing later in life.
I want to relate to the thing that's called a car. The car is for a lot of us, in our childhood, it's a passion. It's pretty cool. At the very basic level, it's a necessity because it takes you places. For some people, it is also the second most expensive item that they ever buy after a house. At the end of the day, this is a pretty sophisticated piece of metal, and nowadays also technology, that is out in the open. It's exposed to the elements and to damage which happens quite frequently.
Apparently, every couple of seconds around the world, some car gets damaged. When you think about the way you interact with cars, you're buying your new car, you're buying a used car. You're filing an insurance claim. When you go abroad, you rent a car, etc. All of these are transactions where cars move, change hands and change liability. We think that there are over one billion transactions like that taking place around the world where a car is changing hands between car rentals. You have subscriptions, car sharing, the old model of leasing and a variety of other transactions. Not to mention new cars that are moved from the production line all the way down through ports and terminals, all the way down to the dealership and to the end customer.
It's a big business and lots of transactions. When you look at how cars are getting inspected now, it's mostly manual. It's someone with a piece of paper. In the best case, there's an app that they use to register any existing damages and create this form, which is problematic because it's very subjective and it's expensive. You have to send the person out there and you have to wait. Most importantly, in many cases, it keeps the car out of the road for a while. Imagine a car that gets hit and waits to be inspected. Imagine a used car that comes back from leasing and instead of marketing it, selling it onwards. There's a great shortage of cars right now, especially in North America. That's quite an interesting topic in its own right but this delay causes inefficiency and it's all because of that manual process of inspection.
It reminds me of one time, I remember bringing the rental car back and they looked at it. Everything's fine and I believe I waited because I usually like to wait but sometimes, I'm impatient, I just run to the terminal. I think this time I've been waiting and we signed off on it. A month later I got a bill and it was $1,000 and it was for some ding that I didn't cause. It took me a lot of phone calls and emails and wasted time. It caused me to not want to have a relationship with that company anymore.
It was one of the two big guys. I can't remember which one now because I've been out of the rental market for a few years. I can definitely see what your tagline and your website say, “Turn your camera into the ultimate vehicle inspector.” It makes perfect sense in the way that it digitizes it. If I would've had that app, you could have seen if there was a dent there when I took it out. When I bring it back, it would've said, “Nope, no dent,” or if it would have been there in the first place, it would have shown obviously.
That’s right. If you look at the angle of the car rental company, I don't think they're bad people. It's that they couldn't keep track of the vehicle damage. It so happened that they track this vehicle right after you used it. If they had this consistent objective view of the vehicle condition, they could have identified who really caused that damage and allocate the damage to the right person.
Thinking about, where does artificial intelligence play a role in your technology?
AI exists in multiple places around this tool. I'm talking about both the mobile inspection solution, where you walk around the car and you take a video and we basically, from there identify the vehicle condition. There is also AI embedded in our stationary camera system, the auto scan where the vehicle drives through a set of CCTV-type cameras and then we create that digital record of damage and whatnot. AI is needed here mainly because we work based off simple cameras, simple equipment, your own mobile phone, any mobile phone or any camera that you have on a roadside. This makes it very difficult for us to control the conditions and control what quality of image we are getting. The first use of AI is quality control. It's to tell you, “You missed a spot. Come closer. It's too dark. Something has gone wrong with this inspection.”
This can only be done when you employ the AI at the edge right there and then, when the person is next to the vehicle. That's quite tricky. There is another layer of AI, which after you've completed your scan, really understands where you are. The vehicle also models the different parts of the vehicle and then starts identifying issues but that's not all because after we identify the issue, what do you do with that? You need to make a decision to repair or replace. How much is it going to cost? How much is the vehicle worth? What's the overall condition of the vehicle? We have a grading system in North America. We need to understand and give the dealer or the buyer that's looking at the car from hundreds of miles away and trying to figure out whether they want to buy the car. We give them tools to understand what this car is worth. This is all data science and it needs to work in conjunction with computer vision and AI.
Now, is it mainly for the body or can you also look at the engine?
It is mainly for physical condition but I’d say it's more than the body because we were using models to understand the overall impact of damage on the vehicle. It is less about faults of, let's say, engine faults and engine codes and things like that. It is more about damage, which inherently comes from the outside, from the usage of the vehicle, how it impacts the vehicle.
Would it be fair to say that if AI didn't exist, then your company would probably not exist or it would be a very different type of company probably?
The economy is incredibly adaptable to new technologies that people doing manual labor would transition to do something different.
Right. That sort of solutions that we've developed over the years that couldn't leverage AI or leveraged it in a minimal way but couldn't make the full inspection objective and convenience, which in turn means that you have to use very controlled settings to conduct the inspection. That puts some doubt on the effectiveness of the solution. The fact that you have AI, that you can control the quality of the images means in practice, that you can put this solution in normal people's hands. You no longer need that professional to come and use this incredible machine. It is so incredible and that is simple enough that almost anyone can use it.
A friend of mine from college is in operations at a company called Carbyne911. It’s a different model with a similar concept. If you see a car accident, theft or anything involving any kind of a crime, they have you connect to your local police station in a live setting and record it, instead of calling 911 with no video. It doesn't make sense. You're on the phone. They can hear it, but now if they could actually see it in a connection that's secure and record it on both sides, it's a pretty neat technology. It’s also powered by AI. I like to ask the elephant in the room question and that is people versus technology. A lot of people say, “This is going to affect my job. There is going to be mass layoff across the country because AI is taking over.” What have you found in terms of people supporting the kinds of things you do? Does it have a negative impact on jobs or is there a positive impact on jobs?
It's a super interesting question and admittedly, not one that I can claim that I've figured out. What I've seen in this field is that layoffs are not the immediate effect. Overall, organizations can adapt themselves and utilize the workforce in different ways. For example, one of the inspection companies that we work with has shifted to offering remote support. The AI actually creates a queue, solves the obvious cases and then transitions them to work on, maybe cases where they need some intervention.
There are some internal issues in the car, somebody must touch the car to assess it. They use the best people to, first of all, support the usage of the technology and intervene in the more complex cases. This is one way. They will probably hire less people and they can grow their business whilst hiring less people than they would without the AI. It’s maybe not about layoffs, but it's about growth with less headcounts. Overall, I think the economy is incredibly adaptable to these technologies and perhaps people over time that would go and do manual labor would transition to do something a bit different. Maybe work less hours, but the overall productivity means that there is enough reward for them too. Maybe that's just an optimistic view of the world but I believe that.
The common theme that I hear is they're redeployed to other areas. The last call that I did was, “I've never had a customer in the hundreds of deployments they’ve done where people were let go.” They already have the budget for the headcount. They make them more productive by focusing on more value-added work. When a computer can do the effort on your site, inspect, auto scan in Ravin Eye, when you see this, you can see it. If a human looked at it, it's not going to see that there's a blemish in certain areas and the 3D image will, 100% of the time. It makes intuitive sense that this is something that AI should handle. What about from a sales motion? We've talked a lot about Ravin, how does your sales team or how do you reach out to people? Are you leveraging any kind of automation or AI in the sales process?
In that sense, I'm embarrassed to say not so much, and we should do more of that. The main reason for that is that we sell to enterprise customers. Relationships do matter a lot. We don't struggle reaching people and making ourselves known. In fact, most of the challenges in the delivery phases, getting the requirements correct and getting the unit project up and running, which we're also getting much better at. In the sales process itself, I think there's a lot more we can do and maybe gathering some data upfront about the needs of that customer. That would be great. As we speak, I'm thinking there is a lot of opportunities there. Not only targeting customers but making the whole sale process more efficient.
Having these conversations with biotech, healthcare, electrical, I had one a couple of days ago with utility, a company that does AI on top of utilities. They build the second world and then they design what would happen. That kind of tree, is it going to touch the wireline at some point, the electrical and start a fire? They can do so much more with less that whenever I talk to all these different companies, bringing it back to sales thinking, knowing that they're doing that with car scanning, how could that apply to a sales motion?
It's interesting looking at all these different industries and bringing them back. The way our customers are using AI is to identify influencers in a network that could make a warm introduction to their end prospect. We worked with this company for about 1.5 years and we loved it. We ended up acquiring the code because if you go through a referral, you're 181 times more likely to sell a deal than if you cold call somebody.
You've experienced that through references. If you look at my entire pipeline, in my sales over the last few years, most of it says Chad's LinkedIn connection and that's a good thing. If you can scale connections like that, that's one of the areas that AI definitely can have an impact. What are you think of the future, whether it's talking about your product, talking about AI in the sales motion, both of those things? Where do you think that the world's headed in terms of leveraging artificial intelligence?
Contrary to some predictions or what we hear on the news, I don't think AI is going to take over the world. I think AI is finding its way to supporting these critical business processes. You described fantastic use cases. I think ours is not bad either. There is probably a handful of other areas where AI can chip in and create efficiency and benefits for humanity. With AI, I struggled to see it doing and maybe it will, is intervening in the more subtle human activities that require good judgment. You're having a conversation with a person on the other side, can you really have a meaningful conversation with AI? I don't think we're there yet.
The emotional intelligence side, the ability to make decisions based on both data but also the relationship with a person. Understanding who that is, what's their set of values and things like that. I think AI will go more into these areas but not in the next several years. Where we will see AI working is identifying the more solvable tasks, removing them from the queue for the people to focus on the more difficult tasks. That is primarily what we are doing. Creating more productivity in the hard industries, manufacturing, medicine, cars, resources and things like that. I think AI can do a lot there.
Eliron, this has been a fabulous conversation. Thank you for coming to the show. You can reach them at Ravin.AI. If you're in the car rental business, if you're in the car sale business, if you're buying a car, if you need to get an understanding of how in shape that car is, you got to check this technology out. It's pretty hot. Eliron, thanks for joining the show. I appreciate your time.
Thank you. I appreciate it.
About Eliron Ekstein
Eliron is Co-founder and Chief Executive Officer. For the past 10 years, he has worked to digitise the automotive and transportation sectors. Before founding Ravin, Eliron was Head of New Business Development at Shell’s Digital Ventures unit in London, investing and incubating new businesses that can leverage Shell’s global footprint of retail location for the benefit of fleets, motorists and smart cities.
Prior to that, Eliron served as the CEO of FarePilot.com, a startup using machine learning to help taxi and Uber drivers find their next passenger by predicting demand patterns in local areas. Eliron has an MBA from London Business School and is a proud father of 2 boys and one adorable girl.