AI for Car Rentals?! Prove It! | Pitch.Tech EP 2

In this Pitch.Tech episode, Rexalto co-founder Loriana Sardar pitches an AI dynamic pricing engine built for the car rental industry, explains why she settled on a flat per-car fee instead of a revenue share, and recounts a Dubai luxury client who called to complain the algorithm's price was too aggressive to believe. Investors John Whaley and TaeHea Nahm probe where the AI actually lives in her model and how she talks about her team. Nahm tells her exactly which slide would have made her traction impossible to ignore.

John Whaley (Inception Studio, Founder): Hi, I'm John Whaley, I'm a three time founder. I graduated from MIT and from Stanford. I won the best thesis award at Stanford, had multiple exits so far. Also started Inception Studio, which is now the top accelerator program for early stage AI founders here in Silicon Valley. I've been an engineer since I was 5 years old, when I started programming. I won the USA Computer Olympiad as one of the top 15 programmers in the US.

As an investor, I'm looking for venture scale opportunities where I can put in a million dollars and have an opportunity for at least 100x return coming back to me within 5 to 7 to 10 years. 

Loriana Sardar (Rexalto, Founder): So we are raising around our pre-seed round of 1.5 million. The goal of today is to meet investors and potentially close our round. Let's go.   

My name is Loriana. I'm a co-founder and the CEO of Rexalto. We have developed and introduced an AI dynamic pricing engine. We previously co-founded one of the largest European car sharing companies. The company generated 75 million in annual revenue and, what's more important, it was cash flow positive.

What's happening now in the car rental industry? Increasing prices of vehicles, uncertain residual values, and a growing percentage of electric vehicles are forcing car rental and mobility companies to put more focus on not just reselling cars as they used to do before, but actually making sure that they make money on their operations. On top of that, for mobility companies especially, they strive to increase their utilization and improve fleet distribution. All of that can be done with the correct pricing.

Taking all of that into account, we have engineered an AI dynamic pricing engine that sets the right price at any given time. First class economics: bringing supply and demand into equilibrium. So the way it works is we look into the company's historical data, market historical data, competitors' prices, and many other factors that help us forecast demand. Things like weather forecast, flight information, booking information. We not only set the price for our customers, but we also push it back either into their fleet management system or to the OTAs, the aggregators where they get rentals from.

We currently focus on North America, Europe, and the Middle East. The segments we look at and work with are either small companies or large enterprise companies. What differentiates us is that we work not only with the car rental industry but with the mobility industry, and the technology, because for mobility companies we use up to 500 data points, and for car rental companies it's about 20, 30, maybe 40.

We offer pricing as a service. For our traction, we currently have 20 paying customers plus piloting customers. We have integrated six reseller solutions. We generated about 200,000 in annual revenue this year. We are raising a round of 1.5 million. We've closed 950, and we're now closing the remaining 550. Thank you very much.

John: You mentioned in the presentation that some of your customers are using this for car rental rates. Are there other things in the automotive space that you're using for that dynamic pricing today?

Loriana: A lot of companies in mobility are using that. Look at Uber, Lyft, other mobility companies. It depends, because you have a lot of smaller players, players for micro mobility for instance, and they use third party solutions. They don't have their proprietary software, and for us they are resellers because we integrate into their platform and they sell to their customers.

John: For any tech-enabled platform, I imagine they're already doing something like this, or some version of it, where they have a pricing engine and can determine prices. This is kind of a common thing. You mentioned Uber and Lyft, they obviously do all of that. But now everything from concert tickets to e-commerce is often doing this type of dynamic pricing.

This might be core to their business, very important, so they may want to develop it themselves. They have all the data about it. What type of customer would use a service like yours versus developing this in-house or using some other solution?

Loriana: Very good question. First I'll answer who are the companies that develop their own. I talked to the head of revenue management of Getaround. They spent a lot of money and many years developing dynamic pricing, and it didn't work. It's not that simple, because you need to know how to do it, not to reduce your revenue even by 1%, because if you do that you're fired. So they said, we cannot just use your service software because we need to make sure our own works, so we'll keep advancing our technology and use it internally.

Answering your question, who are our customers? We have either very small companies that have fifty to a hundred cars. They don't have their own proprietary system; basically they're fleet managers, they're operating their fleet and they want to do it in the most effective way. They have a third party software solution and they compete with budget, enterprise, etc., so they need to make sure that they somehow stand out. These are our customers, and they use our standard, self-service dashboard.

Then we work with large companies, and these companies are not as big as enterprise chains but they have hundreds of thousands of cars, still huge. They basically manage their franchisees, so they need to make sure that their pricing is consistent throughout their network, that they offer the same quality of service, and that their franchisees do well, because they get a percentage of what franchisees earn. So that's our second type of customer: big groups and enterprise companies that manage franchisees.

TaeHea Nahm (Storm Ventures, Managing Director): What makes your dynamic pricing engine better than the competition?

Loriana: First, as I said, it's the number of data points that our engine can absorb and use to make this decision and judgment. You saw that we have AI in the beginning of this.

TaeHea: Every company now flexing Internet.

Loriana: Yeah, but I'll tell you where the AI comes in. So it's a bunch of ML models that work together. This is not AI, but where the AI comes in is that every time our model works in a certain location for a certain customer, it learns from its past performance, and next time makes better predictions, better assumptions, and improves the quality.

John: On that topic, I noticed your pricing model is that you're charging per car. If you're really confident that, hey, this is the ROI we're going to get, why aren't you saying, you're going to have 90% of the upside and we'll just take 10%?

Loriana: I had one customer who did that, and I was telling my team, let's move into that, we'll make sure our customers do better and we'll do better. It didn't work. We still have one customer on that contract and I'm just waiting for the contract to renew to get away from it. The reason why: you always need to have a baseline. So what are you comparing to? Usually you'd compare oranges to oranges, December to December, January to January. But then when things like Easter or Ramadan happen, they move from one month to another.

Then they'll say, yeah, but the increase wasn't because of your pricing model, it was because Easter was in this month and last year it was in a different month. So much time is spent on this argument, because they don't want to pay you more, they want to pay you less. We said, at this point we cannot do this, it's crazy. So we said we want to be clear and straightforward. We'll do standard pricing per car per month, and that's good enough.

John: But that also points to the fundamental problem of measuring the improvement, the lift, that you're going to have from this. How do you compare apples to apples when you're looking at over time.

Loriana: ...revenue over time? They have dashboards, they see how the pricing increased over a longer period of time. Maybe one month to another might be tricky, but if you look on a yearly scale, you still see the increase.

TaeHea: Do your customers do A/B testing? Can they run two side by side?

Loriana: We usually start with A/B testing, especially with big customers, but we don't do it within one city, because you cannot do it in one location, the customer will choose what is cheaper. So you have to do it in different locations. We would choose a few selected locations. Usually the way our pilot works: two or three locations. One would be maybe the top location, another would be a mid-sized location, and the third would be a small location. And then we see the results in those.

TaeHea: And what has been the result?

Loriana: Our worst result was plus 5% revenue increase. Our best was six times, but that was a very particular case: a luxury car rental company that was staying within its boundaries.

We analyzed the supply and said, it's Dubai, end of December, there are two Lamborghinis available and there is demand for 100. You can put whatever price you want, you will still sell it.

TaeHea: That doesn't sound like an AI recommendation.

Loriana: No, I mean, they called us, they called me, they were like, your system is crazy, it's telling me to put... I don't remember the price, something is wrong. I was like, hold on, let me look into this. So I looked into it, and we said, that's what you should put.

John: Customers don't like dynamic pricing, because it's like, wait a second, you got a different price than I got when I went for this. There's a lot of backlash around surge pricing, these sorts of things, which ultimately ends up creating brand risk. So even though companies could potentially do this, they're careful about when to do it. What are your thoughts on that?

Loriana: First of all, revenue is not the only KPI our customers use us for. In many cases they want to avoid human mistake. Most of them look at competitors' prices and do it manually, and the level of complexity, you have different zip codes, cars divided into codes, different lengths of rental (one day, two days, three days, etc.), different locations. So the number of price points you have to change every day, sometimes even every hour, is tremendous.

If you do it all manually, you make mistakes, and this is what our top managers tell us: we know this isn't ideal, but we often have to cut corners, we can't put too much effort here because there's too much, so we make an assumption without basing it on data. So automating the process and pushing it back into the system is the number one KPI for many of our large companies.

Secondly, we don't change the price depending on the individual customer. That's up to the company, whether they want different customers to have loyalty programs. For mobility companies we do; we don't look at a particular customer because we don't have that data, but we cluster them, so we understand that a customer falls into a certain cluster, and there's a different coefficient. That's what mobility companies are known for, because they give the car to different audiences.

John: This is very interesting, because I went to your website, and it's a very nice-looking website, but to be honest, looking at it I had no clue what you guys did. Hearing the pitch, I now have a much better idea of what you're doing. At least in terms of the copy on the website, it didn't come across.

You cover the team extremely early, which, to be honest, before I know what problem you're solving, it's hard to judge whether this is the right team. So sometimes it's better to say that later: oh, and by the way, here's the evidence for why we're the right team to do this. So, what is your connection to this problem, this space, and how about the rest of the team? Why are you uniquely suited to be the ones in the world to solve this problem?

Loriana: Actually, the reason I showed it in the beginning is that in our past company, when we exited and decided to do something else together, we started thinking, let's share our knowledge with other companies. We definitely decided not to build another car sharing company anywhere in the world, it's a very operationally intense business. We thought maybe we'd do consultancy.

Then we said, let's do what we did best, and we were really the best on the market in terms of pricing, correctly distributing the fleet, being considered the low coster, and yet making more revenue than our competitors. So we said, let's try to replicate it for other companies. We started initially with car sharing and mobility companies, but then we said, there's a much bigger market here: car rental, which is a dinosaur industry. So we started looking into that, talking to potential customers, and that's how it all started. Thank you very much.

I grew up in a very international environment. I'm half Russian, half Indian, so I was always exposed to different cultures. I graduated from two US universities, both in Russia, then I traveled abroad. I started my previous company together with my co-founders back in 2014, and we ran the business for eight years. I left the country in 2022 for the war, and it was one of the most difficult moments of my life, very stressful, very uncertain.

We came to the US together with my family, with my husband. After about half a year, I said I want to build something new here. We wanted to use the knowledge and experience we gained from our previous company, and the understanding of the market, and share it with other companies, not only car sharing companies but mobility companies and car rental companies too.

TaeHea: The thing I like about Rexalto is that she's a successful founder, that she built a very successful business before, and based on that experience has come up with a solution to this problem in a compelling manner. What would be great for me is fairly early on, once she explains her product, to have a slide that says: I have six customers right now. In the beginning I'm wondering, is this an idea?

Does she have customers? Does it work? But if I know she's got six customers who like it, then I know this is great. And then the next thing on the same slide is to show: in one year I'll have 20 customers, and these are the customers I'm working on. And then maybe five years from now, this is what I want: 1,000 customers.

John: All the pitches talk about AI. To be honest, many of these are more of a buzzword, they're using it more as a buzzword. I know we need to talk about AI, so let's try to craft an AI story onto our business. They didn't really talk about why they're using AI, why it's important.

Loriana: Why do we have a tech domain? I think the domain you choose should reflect what you're doing. I like the idea of highlighting that it's a tech company, a tech product. And I think we stand out from our competitors because we highlight this point.