Silicon Valley has a comfortable office with sparkling meeting rooms and on-site baristas. So when Happyrobot engineer Ari Polakof decided to start by attacking his company, he didn’t expect exactly to slap his laptop in the service bay along with the mechanic.
“It’s so loud, it’s impossible to concentrate,” he laughed.
But that’s part of the story of how Polakof’s new Car-for-Car-Dealerships startup got off the ground last year.
Founded with Y Combinator’s Polakof and his brother Alen (also from Happyrobot), and former Netflix data scientist Juan Alzugaray, Flai is one of many startups seeking to use artificial intelligence to improve the experience of purchasing, selling and serving cars at dealerships.
The co-founders built software from scratch, specifically tailored for the automotive dealer environment. As Polakov said in his interview, it is “omnichannel.” This means that you can handle phone calls (using voice agents) and emails and text (using a larger language model).
Now, Flai is closing its $4.5 million seed round and expanding what they’ve made. The round was led by Liz Wessel in the first round capital and included funding from YC, Red Blue Capital, Joe Montana’s Liquid 2 venture and innovation efforts.
Flies aren’t the only ones doing this kind of work. Earlier this year, another YC-backed startup called Toma announced that it had raised $17 million from automotive industry influencers such as A16Z and Yossi Levi, known as the automotive dealership guy. While other packs are trying to build similar products, legacy interactive voice response companies (the companies that make phone tree software that we are all familiar with) are trying to keep up the pace.
TechCrunch Events
San Francisco
|
October 27th-29th, 2025
CEO Polakof said the Flai offering differs in that it does not use off-the-shelf audio technology. The startup essentially built everything from scratch. In particular, Polakof claims he has already persuaded several dealers to switch from other companies, as he made the voice agent impressive.
What’s more, he said there’s a lot of room for competition in the market. There are thousands of car dealers and service centers in the United States, and most of them have the same problem. If your phone line is tied up, you could lose potential customers.
After getting off the ground last year, the FLAI team was to train AI, realize the best way to acquire customers and start appearing at those dealers.
Sometimes they would appear in cold calls, emails, or LinkedIn messaging, or in network meetings. However, most of the early work that Flai completed was through direct visits to around 400 brick and mortar stores in the early days, due to Polakof’s count.
When Flai attacked these early relationships, Polakof said it was essentially embedded in each dealer. That’s how he realized he was in Service Bay last year, but he means that more often it means setting up in an empty office.
Polakof said the Flai team spent “every day, all day” during that first period. “It’s almost very tired and painful, but I think that’s the only way to do it. And I can’t imagine a big company doing it,” he said.
As startups move from that learning period to something focused on making customers happy, Polakof said, like many of their Silicon Valley peers, the smaller team is still pushing for a long time.
Seed’s funding will help FLAI grow, but Polakof said he doesn’t think the team will explode.
“We want to work as smart as possible,” he said. “I don’t want 100 or 200 people working on this right now.” Instead, he said that the team of three “holds accountable” one another by focusing on the tasks that “100% required, otherwise worthless.”
In that sense, Polakof is not far from his early days working in the service bay or in the dealer back office.
