Redcar's AI sales superpower platform lands $5.3M from Khosla, Greylock and others

The startup's DNA in consumer products like Google Assistant help it combine the best AI has to offer with a UX that salespeople actually like

Redcar founding team, including (left to right) Cataleya Jiang, Ben Wu, Jeff Chen and Aidan Lynott

Sales enablement and go-to-market process augmentation has been one of the early beachheads for B2B AI tools, and Redcar, a two-year old startup based out of SF, fits squarely in that category. The company is building AI sales agents that take care of a lot of the repetitive, fundamental steps involved in the sales process so that the people doing the selling can actually focus on the close.

The company revealed $5.3 million in funding today, including a $4.3 million seed round led by Khosla Ventures and including participation from Greylock, and a $1 million pre-seed from Humbition, as well as individual checks from some heavyweights including Mitch Kapor, Siqi Chen, Nancy Xu and Elias Torres. It’s a remarkable list and a sizeable early financing, in a crowded category, but speaking to two of Redcar’s co-founders – CEO Jeff Chen and Head of Growth ‘CJ’ Cataleya Jiang, it’s clear why they were able to stand out from the competition.

Chen is a serial founder whose previous company Joyride built Android’s top voice assistant, Skyvi, and which exited to Google in 2015. He launched Google Assistant with the company, then eventually moved on to build another startup, Taste, which was a part of Y Combinator’s winter 2021 batch. It was at YC that Chen really dug into the sales process.

“One of the rites of passage for founders is how hardcore you are,” Chen said. “And you demonstrate this ‘hardcoreness’ by showing how many cold personalized emails you can send per day[…] I wrote 200 a day.”

After a while of doing that as part of a two-person team for Taste (with Redcar co-founder and Chief Software Architect Ben Wu), Chen decided it might be better to outsource it, and contracted someone to handle it for him for $5K per month. That ended up netting only one meeting with a potential customer, so he rolled up his sleeves and took over again himself, landing about 50 such conversations. It wasn’t sustainable, but something just over the horizon would make it manageable – as well as scalable.

That was the arrival of the large language model, which instantly struck Chen and Wu as something that would provide a much better and more transformative opportunity than what they were originally working on at YC. Redcar grew out of Chen’s experience with trying to stock the funnel for their prior company, combined with the efficiencies that LLM-based AI brought to the table in terms of handling a lot of the manual lift involved in that process.

Redcar helps B2B companies scale sales teams by providing building blocks to create tailored AI agents, letting human reps focus on closing deals. There are three core components to the product: custom research agents, prospecting, and personalized outreach. Part of the startup’s unique value prop is the degree to which these components can be customized to the customer, and to the sales rep, and combined in different ways to best match their workflow and particular sales motion.

Chen and CJ describe the approach as a ‘verizontal’ one, which describes the way in which packaging and coordinating a few vertical tools aimed at specific functions within the sales process (prospect research, for instance) ends up providing a much more broad, horizontal solution.

"If you combine a few vertical tools together, it starts becoming like a bundle,” Chen said. “That's what we think is a 'verizontal.' The better AI gets, the more we can make more software, and the more bundling happens.”

It’s an approach that could be (and is being) applied to a number of different problem areas, so I asked Chen why they started with sales as the immediate focus for Redcar. Chen explained that in addition to his own experience and frustration with the sale process, in part, it’s because of how the current state of the technology fits the shape of the challenge.

"Language models are probabilistic by nature,” he said. “You want to be in a space where you don't have to get things perfectly right – where 99.9% is good enough."

As for why CJ is interested in tackling sales and GTM specifically, she said that it’s surprisingly similar in some ways to her background studying theoretical mathematics in school.

“It's very logical, I really enjoy it,” she said. “It's almost like a puzzle piece you solve – you gather all the information and then it’s putting the blocks together.”

Creating and assembling this blocks is exactly what Redcar is all about, but Chen is clear that it isn’t about replacing the human elements of the sales process yet. He sees it as upleveling the profession in a way that has ample historical precedent.

“In the sixties, we had human computers [...] mostly women of color who did tedious mathematics by hand,” he noted. “When machine computers got invented, those same women became scientists and engineers – now we can do the same for our generation.”

Like many AI builders I’ve spoken to, Chen is partial to the analogy of his technology acting as an Iron Man suit for the human operator inside – it’s a technology that can take any ordinary person and put them on a level with a superhuman peer or adversary. Tony Stark without the suit isn’t going to be able to go toe-to-toe with Thanos, but the suit can make it so that anyone can stand a chance. Redcar, similarly, aims to make every sales person a top performer – all while letting them focus on the most rewarding part of the job.

"Closing is fun,” Chen said. “Who doesn't like to win a deal? But to get to that close... you have to persistently follow up, make your pitch deck, talk to the person. We're basically setting these people up for success so they can enjoy the fun parts."