This is the first large-scale deployment of ZeroToOne.AI’s Large Behavioral Model through GroundTruth’s enterprise and SMB customer base, bringing true behavioral prediction to advertising.
NEW YORK, May 28, 2026 /PRNewswire/ — GroundTruth, a ZeroToOne.AI company and one of the world’s largest real-world signal platforms, today announced the launch of Dynamic Intent Prediction, the first large-scale commercial deployment of ZeroToOne.AI’s Large Behavioral Model (LBM). A new class of AI-powered ad audiences that predict who is likely to act, what they are likely to do, and when they are most likely to do it, Dynamic Intent Prediction is built on the same foundation model architecture that powers ZeroToOne.AI’s enterprise product.
Powered by ZeroToOne.AI’s patent-pending Large Behavioral Model (LBM) developed by AI researchers recruited from Carnegie Mellon University, these audiences refresh every 24 hours and predict in-market behavior across windows ranging from 24 hours to 30 days.
The Large Behavioral Model is similar in architecture to a Large Language Model, but trained to predict future human behavior rather than the next word. It ingests 15,000+ behavioral signals across 7M+ U.S. points of interest, retrains weekly, and scores 2B+ MAIDs across 9,000+ categories and brands for forward-looking purchase and visit intent.
Dynamic Intent Prediction identifies who will convert and when, not who converted before. Most audiences sold as “AI-powered” or “predictive” today are just lookalikes: refined historical segments built on past behavior that produce a static list of people who might convert someday. That’s guessing, not predicting. Every 24 hours, the model updates predictions at the id level, identifying consumers who are in-market right now.
Dynamic Intent Prediction audiences delivered a footprint nearly 50x larger than standard audiences, reaching the vast majority of the targetable consumer universe with daily-refreshed predictive AI segments. Its highest-confidence predictions are 8-9x more likely to convert than a standard audience, meaning less wasted spend and more budget going where it actually works.
“The industry has been describing every refinement of historical data as ‘predictive.’ It isn’t. With Dynamic Intent Prediction, we’re giving advertisers audiences that tell you who will convert and when at a scale nothing else in market can match. This is the first of many innovations coming from the GroundTruth and ZeroToOne.AI combination.” said Rosie O’Meara, CEO, GroundTruth.
Predictive AI audiences spanning QSR, Retail, Automotive, Hospitality, and more are available immediately to all GroundTruth customers, through managed services for media agencies and enterprise brands, as well as media partners and SMBs in Ads Manager, GroundTruth’s self-serve omnichannel ad platform. Audiences can be activated on all media channels: CTV, mobile, streaming audio and podcasts, digital out-of-home or desktop.
ZeroToOne.AI’s LBM has already been deployed across multiple use cases for a global restaurant chain, spanning competitive conquest, retention and upsell, offer management, CRM lifecycle management, and app growth. The results: $100 million in wasted media spend identified, a 27% increase in store visits, and a 3x increase in app downloads while improving media efficiency by 30%.
A leading global automotive brand deployed ZeroToOne.AI’s LBM across media efficiency and competitive conquest, then expanded to regional dealer activation and financial services. Results included a 30% increase in media efficiency, a 400% increase in website traffic, and a 10% increase in financial services revenue.
“Predicting human behavior is one of the hardest problems in AI, and the most consequential. ZeroToOne.AI’s Large Behavioral Model was built to solve it, not just for advertising, but for any enterprise decision that depends on understanding what people will do next. Dynamic Intent Prediction is the first product bringing that capability to market at scale. It’s the beginning of a much larger platform.” – Naseer Hashim, Co-Founder and CEO, ZeroToOne.AI.
Dynamic Intent Prediction audiences are available now to all GroundTruth customers.
About GroundTruth
GroundTruth operates a scaled real-world network that captures and processes billions of daily interactions across the physical world. Leveraging observed consumer behavior, GroundTruth provides high-fidelity, privacy-conscious data infrastructure that powers predictive models and enables a deeper understanding of real-world behavior.
About ZeroToOne.AI
ZeroToOne.AI (ZTO) is an enterprise AI platform built to predict real-world behavior before it happens, enabling brands, advertisers, and enterprises to move from historical analytics to forward-looking decision making. At the core of ZTO is its patent-pending Large Behavioral Model (LBM), developed by AI researchers recruited from Carnegie Mellon University, the nation’s top-ranked AI program. ZeroToOne.AI helps enterprises act on intent rather than past performance.
Media contact
Ander Mateos – Ander.mateos@groundtruth.com
SOURCE GroundTruth