Hyperplane, a San Francisco-based startup building foundational models to help banks predict customer behavior, today raised $6 million in funding led by former Stripe executive Lachy Groom. Ventures, Liquid2 Ventures, Soma Capital, Latitud, Atman Capital, Crestone VC, and Norte announced the round and came out of stealth with participation from Clocktower Technology’s SV Angel. The general idea here is to allow banks to use their data to predict user behavior and build personalized experiences.

The company already has partnerships with about a dozen banks in Brazil, and is now looking to expand into the United States. Hyperplane is currently focused only on the banking industry, but over time, the team plans to bring its technology to other sectors as well.

Hyperplane was co-founded by Felipe Ramunier, Daniel Silva, Rohan Ramanas, and Felipe Meneses.Ramnier (CEO) has spent the past seven years start setis a Brazilian EdTech startup whose members Daniel Silva and Rohan Ramanas previously built large-scale AI systems at Google and LinkedIn.

“The core hypothesis we started with was: What does it take to build a layer of personalization for banks around the world?” Ramanath explained. “If you think about big tech companies, they have a lot of first-party data, but in order to use all of this data to understand the consumer and build personalization, they have to rely on their data infrastructure and enterprise data. We’re also investing heavily in warehousing, where we create every product page and ultimately incorporate this into the consumer experience itself. Hyperplane’s goal is to help banks around the world store large amounts of first-party data. , what does it take to build a data intelligence layer so banks can connect their first-party data?”

Lamnier also highlighted the fact that banks have detailed data about their customers that is not available with other services. “One of the arguments I often use when pitching banks is that the data these banks have about me as a customer can tell me much more about my behavior than the data Google or Facebook has. You’re vulnerable. Visiting Porsche’s website doesn’t mean you can buy a Porsche. But Chase and Bank of America can’t tell you what kind of restaurant I go to or what grocery store I go to. How much does it cost to take Uber? We have all that data in-house.”

Currently, most banks offer little personalized experience, so the baseline is low. But consumers increasingly expect their banking experience to be similar to other online experiences, especially in competitive banking markets like Brazil. At its core, Hyperplane provides banks with an API to build these personalization models on the fly. The team stressed that all of these deployments are private and no data sharing will take place. Hyperplane also uses its own models for all of this.

Currently, the company offers two modules. One is for building audience segments and the other is for creating lookalike audiences to find similar users and expand your potential target audience. ” We find that by building task-specific models, we can get more benefits from our construction. It was something custom and made from scratch,” Ramanath said.

Most recently, Hyperplane launched the Mandelbrot LLM. This is a particularly useful model for predicting when banks will churn customers and which users will treat a particular bank as their primary bank.

Hyperplane says that by using its services, the credit line division of a Brazilian neobank, for example, has been able to get a clearer picture of its customers’ estimated income, increasing transaction volumes by 46%.

“Brazil has experienced an important pro-competitive movement over the past decade, and we now see an ecosystem eager to adopt new technologies,” Ramnier said. “Hyperplane’s cloud can be scaled across markets with little effort. We will be announcing our first partnership in the US soon.”