Nvidia's New AI Predicts Customer Churn Without Any Training
This could save businesses weeks of work and millions in lost revenue.
NVIDIA introduced Kumo Relational, a pre-trained relational foundation model that makes predictions directly on enterprise databases. Given a predictive query, it gathers up to 10,000 historical examples as context and predicts outcomes such as churn, demand, or fraud in a single forward pass—with no task-specific training, no feature engineering, and no flattening of tables.
According to the article, it is the first few-shot foundation model to outperform supervised machine learning on common relational benchmarks: across 41 tasks on 15 databases, it beats supervised relational deep learning models and expert-engineered gradient-boosted tree pipelines, and on the SAP SALT enterprise benchmark it surpasses AutoGluon ensembles by 8% in mean reciprocal rank. It scales to databases with 500B+ rows and works for humans and agents alike through NVIDIA’s open-source, GPU-native structured-data-models library or as an NVIDIA NIM.
- Nvidia's Kumo Relational predicts business outcomes like churn or fraud without any training, saving weeks of work.
- It outperforms traditional machine learning on benchmarks, beating XGBoost by up to 13 points.
- It's built for large enterprises with massive databases and can be used by AI agents for automated decision-making.
Why It Matters
Businesses can make faster, data-driven decisions without hiring expensive data scientists, potentially saving millions.