The CEO and co-founder of Pecan AI, Zohar Bronfman, explained that generative AI models are not good at transforming raw tabular data from different sources into the format required for predictive modeling, a skill that typically requires an experienced data scientist. To address this, Pecan AI is working on automating data preparation and feature engineering, including identifying data leakage, which can undermine model accuracy. The company has raised $116 million in funding since its inception, including a $66 million round in February 2022.
Key takeaways:
- Pecan AI, an eight-year-old startup, is launching a new tool, Predictive GenAI, which combines generative AI capabilities with predictive machine learning.
- The company aims to democratize AI capabilities and make them more accessible to business users, not just data scientists.
- Pecan AI's Predictive GenAI has two main features: Predictive Chat, which allows users to make natural language queries, and Predictive Notebook, which uses generative AI to build the data science notebook for predictive modeling.
- Pecan AI is also working on automating data preparation and feature engineering to improve prediction, including addressing issues like data leakage which can undermine model accuracy.