Dataminr Launches ReGenAI: Combining Predictive and Generative AI for Real-Time Information Insights

Modern generative AI relies heavily on user-inputted prompts to deliver results. But what if there were a system capable of predicting the optimal prompt to generate the best outcomes?

Introducing ReGenAI by Dataminr, a groundbreaking approach that transcends traditional generative AI.

Are You Ready for AI Agents?

Dataminr has been enhancing its real-time AI platform since 2009, long before generative AI surged into the mainstream. Over the past decade, the company's technology has evolved to help organizations like the United Nations, Airbus, and Dyson assess potential risks by analyzing data from various sources.

While Dataminr previously used generative AI to provide text summaries of detected events, ReGenAI elevates this capability. Now, the system can regenerate event alerts in real time as circumstances evolve.

“This is the next evolution for Dataminr,” says Ted Bailey, the company's Founder and CEO.

Moving Beyond Prompts

Bailey emphasizes that awareness of an event is powerful, but corporations and governments need insights into its scope and potential implications. ReGenAI dynamically detects new developments from diverse data sources, synthesizing real-time information and updating recommendations as events unfold.

“This use of generative AI transcends prompts; it aims to automatically reformulate based on real-world connections,” Bailey explains.

Predictive and Generative AI Combined

Many modern technologies, including generative AI tools, suggest next steps for users. However, Bailey clarifies that ReGenAI offers something different. It isn't merely an agent-based approach linking multiple generative AI models; it's a more sophisticated system.

Dataminr has developed its own large language models (LLMs), trained on a decade's worth of data from historical events. “Today, all our models are internal, thanks to a mix of customized and open-source training,” Bailey shares.

When an event is detected, Dataminr's predictive AI models create a chain of related insights using a variety of data sources. “As we process streaming real-time data, we uncover interim developments and describe them using generative AI,” he notes.

ReGenAI builds on existing predictive and generative AI models, applying real-time logic to help users grasp the significance of events at any moment. “It’s the fusion of classic predictive AI and generative AI, resulting in dynamic, iterative live information,” Bailey concludes.