Hyperplane

Hyperplane develops foundation models for banks to anticipate customer behavior by leveraging their first-party data. Its LLM, Mandelbrot, is designed to predict customer churn and identify primary bank users.

Available through an API, Mandelbrot embeddings can be fine-tuned to train predictive models using institution-specific data. This allows for the creation of nuanced customer personas and marketing campaigns based on financial behaviors and consumer interests while providing regularly updated user income estimates to expand user limits without concern for defaults.

Hyperplane collaborated with several banks in Brazil, with plans to expand its footprint into the US market as reported in December 2023.

The platform offers modules such as:

(1) Audience Segmentation: Allows users to segment their customer base using various criteria, facilitating targeted marketing campaigns, content delivery, and user analysis.

(2) AutoML - Lookalike Audience: Enables the expansion of target audiences by identifying similar users from a seed set, leveraging Hyperplane's Knowledge Graph and Foundation Models for model training and actionable insights.

(3) Income Estimation: Provides tools for estimating individual or household incomes based on transaction data, facilitating accurate income estimation and verification through processing transaction data and other relevant inputs.