While most AI startups are busy building bigger language models, SQREEM is heading the other way. The company’s new CEO, Stephen Yap, is making a clear bet: marketers need to know what people do, not just what they type or say. SQREEM’s large behavioral model, or LBM, is built to find those hidden patterns. Yap thinks that’s the edge brands really want.
Reuters recently reported that major AI companies are under increasing pressure to prove the value, safety, and differentiation of their models, with Anthropic and Accenture announcing a $2 billion partnership to evaluate frontier AI models over five years.
Instead of reading language, SQREEM’s LBM looks at real-time behavioral data from social platforms and open web sources. The model builds audience personas by mapping links between actions-like the odd connection between green tea buyers and people shopping for travel insurance. Founder René Raiss says the system can follow a person’s path from searching for green tea to looking up antioxidants, then skin care, sunscreen, and finally vacation planning. He claims this gives a much deeper view of what people want, far beyond basic demographic targeting.
Reuters has highlighted a broader trend in the AI sector, where companies are increasingly using behavioral signals alongside self-reported data to improve targeting and safety. For example, a spokesperson for OpenAI told Reuters that the company uses both account and behavioral signals in addition to self-reported age, illustrating how behavior-based inference is becoming standard practice among leading AI firms.
For marketers, the pitch is simple: instead of guessing what people care about, the LBM claims to show the real questions and values behind buying decisions. Yap says this helps brands write messages that actually connect, instead of relying on generic demographic data or keyword lists. The model can spot “numeric entanglements”-statistical patterns that point to likely outcomes-though Raiss admits it can’t predict pure chance or once-in-a-lifetime events.
All this depends on data that, according to SQREEM, is pulled in real time but never linked to individuals. The platform tracks public mood and behavior shifts before and after big events. It runs sentiment analysis on social comments to measure not just how much people are talking, but what they’re actually saying. Brands can plug SQREEM’s tools straight into their dashboards, making behavioral insights part of their daily work.
The AI field is crowded with companies promising the next big thing. SQREEM stands out by focusing on what people do, not just what they say or write. This is similar to the AI audio shift that has pushed publishers to rethink how they reach and make money from audiences.
Marketers and publishers tired of chasing the latest AI trend may see SQREEM’s approach as a practical change. The company’s choice to skip the LLM race and focus on behavior could give brands a clearer view of their audiences-if the model’s claims hold up. In a market where every tool promises to predict the future, SQREEM is betting that knowing what’s happening now is the real advantage.
