San Francisco-based Onton has unveiled Ontology 1, a new AI model engineered for trustworthy product discovery on the emerging agentic web. This model addresses a critical problem: AI systems increasingly research and execute consumer purchasing decisions, but often rely on manipulated or synthetic information.
Built from scratch, Ontology 1 aims to cut through the noise of incentivized recommendations and fake content. Onton's co-founder, Alex Gunnarson, stated, "Everyone is focused on building smarter agents. We're focused on a different question: what should those agents trust?"
Outperforming Existing Discovery Platforms
Onton claims its model outperforms dominant players like Google Shopping and Amazon in accuracy benchmarks. The performance gap is most pronounced in verifying product information veracity, a domain historically difficult for AI.
For example, an AI agent tasked with finding a sofa under $2,000 might sift through thousands of results, reviews, and influencer posts. Without a reliable way to discern signal from noise, the agent risks recommending products based on untrustworthy data. Onton was built to solve this problem, which the company views as an escalating default state of the internet.
A New Foundation for Agentic Commerce
The modern internet was designed for human evaluation, where intuition and skepticism filtered content. As AI agents take over more decision-making, these human filters disappear, leaving systems vulnerable to manipulation.
The daily brief
Get the free daily brief with the day's AI news, new AI startups and funding rounds every morning.
Join 64,000 other subscribers.
Onton's model evaluates not just what a product is, but the trustworthiness of its surrounding information. It interprets user preferences, including visual inputs, as meaningful data rather than potential noise. This ensures AI agents make recommendations grounded in reliable information. The company also released new research detailing how current AI systems fail when confronted with synthetic content and incentivized recommendations, concluding that existing systems are ill-equipped for this information environment.
StartupHub.ai data shows Onton, with a score of 54/100, is still developing its market presence compared to established competitors like Gong (76/100) and Five9 (62/100). Onton's co-founder Zach Hudson emphasized, "The next major internet platform will not be defined solely by who has the best model. It will be defined by who can provide the most trustworthy foundation for those models to operate on."
Onton views a trust layer for agentic commerce as a precondition, not a feature. Without it, every AI-powered purchasing decision risks being built on a gamed foundation. With an agentic web authenticity model like Ontology 1, agents can fulfill their promise: helping people make better, faster decisions with reliable information. This approach is critical for trustworthy product discovery AI in a world where skills are the new SDKs for AI agents.
Ontology 1 is available today via Onton.com and for partners building on the agentic web requiring a trustworthy foundation for product discovery and recommendation.
