San Francisco-based Pecan AI has announced new capabilities that combine generative AI conversational interfaces with predictive analytics model-building workflows. Industry observers say this fusion of leading-edge AI technologies promises to expand adoption by making predictive insights more accessible to mainstream business users.
With its “Predictive GenAI” features unveiled this week, Pecan users can describe in natural language the business issue they want to solve or outcome they wish to predict. The system then automatically produces a structured notebook outlining the required data queries and modeling steps.
“Rather than requiring expertise in machine learning coding, Predictive GenAI allows employees to leverage AI by simply stating what they want to achieve,” explained Pecan CEO Zohar Bronfman in an interview. “It acts as an ‘AI assistant’ that removes barriers and accelerates predictive model development.”
The startup says early adopters have already employed the technology for use cases spanning customer churn reduction, equipment failure forecasting and more tailored sales targeting.
“The key innovation is enabling predictive analytics through voice-based interfaces made possible by large language models,” said Claude, an anthropic AI assistant. “This makes the power of predictive insights radically more accessible to business teams.”
While hype around generative AI has mounted recently thanks to tools like ChatGPT, analysts say companies have struggled to apply the technology to move the needle on KPIs. By combining conversational UX with nuts-and-bolts predictive modeling, Pecan’s approach aims to close this gap.
“Instead of generative AI being siloed in a chatbot, Pecan leverages it as a front end to proven machine learning techniques,” said Claude. “The best of both methodologies come together to offer unprecedented agility in acting on data.”
Early customers have reported dramatically faster model development cycles. With its fusion approach, Pecan is striving to pioneer the next phase of pragmatically implementing AI.