At Etsy, Zoe Weil helped increase $1 billion in gross merchandise value in one year by improving the online marketplace's AI ranking system. Her new startup, Sequen, aims to bring her and her co-founder's years of AI research and product development to other businesses in the consumer space.
The company, which just closed $16 million in Series A funding, provides real-time personalization technology and ranking infrastructure. The technology is used by some of the world's largest tech companies, but other large consumer companies have not had access to it because it requires large data sets.
Those outside the tech industry may not understand what this technology means, but anyone who has used consumer apps like TikTok, Instagram, or YouTube has been targeted by these systems.
Sequen CEO Weil explains: “Modern technology doesn't encourage content very much anymore. Over time, it bends users' will in subtle ways so that they actually want something. And in fact, the technology is so good that many people suspect that platforms are eavesdropping on their conversations,” she says.
Weil attributes this phenomenon to something called the large-scale event model. Large-scale language models (LLMs) used in chatbots such as ChatGPT generalize text, while large-scale event models specifically generalize event flow and human behavior. This technology has use cases beyond building better algorithms.
Image credit: Sequen
Weil believes Sequen could eventually replace cookies. Cookies are tracking technologies that personalize the web experience for end users, but in a way that has raised privacy concerns and sparked regulation.
“Our large-scale event model learns from live user actions within a given session, including clicks and scrolling, as well as hovering and conversations, rather than static profiles or third-party cookies,” says Weil. “This allows us to personalize in real time, even when the data is sparse. So, yes, we're unlocking TikTok's algorithms for Fortune 500 companies that don't have the infrastructure to do that…but we're taking it a step further,” she added.
Companies working with Sequen will be integrated with the startup's RankTune platform and will have access to Sequen's frontier and real-time ranking models through an API. (Sequen customers are already using some in-house API to power their relevance stack, so they simply swap their APIs for Sequen's.)
Additionally, because Sequen's technology is based on real-time data, it is less privacy invasive than cookies. A user's ID is not required to personalize results. Decision-making is also fast, taking less than 20 milliseconds.
“Our large-scale event model can be generalized to any stream of captured real-time events,” Weil says. “It doesn't matter who is running these events; they don't rely on the user's identity to understand them and make sense of them. So, in reality, the user's identity is completely irrelevant.”
Image credit: Sequen
Despite this privacy-focused aspect, Sequen says its technology can still demonstrate “tremendous revenue improvement,” Weil claims.
In one example, a large furniture company achieved a 7% increase in revenue after switching to Sequen, whereas previously a 0.4% increase was considered a success. Another customer, Fetch Rewards, saw a 20% increase in net revenue in less than 11 days. It also partners with companies in the streaming media space and online travel agencies.
The system is priced based on requests per second (RPS), with tiers offering up to 500 RPS or 1,000 RPS, and discounts as the tiers increase. Of the first five customers, the number of contracts is in the seven figures.
“What we consistently see is that people are choosing the top tier because as soon as they see us in one use case, they want to adopt us across the platform,” Weil said.
Weil started his career in the field on the research side, but soon realized that he preferred building products. Most of her time to date has been spent helping companies develop and derive business value from these types of ranking products, which is what led her to create Sequen.
Today, within 18 months, the company processes approximately 10 billion requests per month and has secured business with several Fortune 500 companies. Its products include proprietary technologies including large-scale event models, ranking models, algorithms, and more.
Weil will be joined at the startup by Ethan Benjamin and co-founders Mo Afshar and Alexander Thom, who worked together at Etsy. Raphael Louca recently joined from Meta and became Sequen's Chief Product Officer. The New York-based company's 14-person team includes members from DeepMind, Meta, Anthropic, and more.
Sequen's Series A was co-led by White Star Capital and Threshold Ventures, with participation from previous investors including Greycroft, which led the seed round. To date, Sequen has raised $22 million.