Click here to get this post in PDF
The San Francisco company unveils proprietary research and a from-scratch model designed to help AI systems cut through synthetic content and manipulated recommendations as purchasing decisions increasingly shift from humans to agents.
SAN FRANCISCO, Calif. – July 29, 2026; Onton today announced the launch of Ontology 1, a trust and authenticity model for the emerging agentic web, where AI systems increasingly research, recommend, and execute consumer purchasing decisions on behalf of real people.
The model was built from scratch. In head-to-head benchmarks against the most dominant players in product discovery, including Google Shopping and Amazon, Onton’s model outperformed on accuracy across every dimension tested. The gap is sharpest in a domain where accuracy has historically been hardest to achieve: interpreting the veracity of product information.
Consider what happens when a consumer asks an AI agent to find the best sofa under $2,000. The agent searches. It finds thousands of results and proceeds to review aggregations, influencer recommendations, sponsored comparisons. However, the risk is that much of this data is synthetic, incentivized, or manipulated. The agent has no reliable way to distinguish signal from noise, and proceeds to recommend anyway. The consumer buys based on information that was never trustworthy to begin with.
Nobody flags this as a failure, since the result simply looks like a bad purchase. This is the problem Onton was built to solve. What was once perceived as an edge case is now increasingly the default state of the internet.
“Everyone is focused on building smarter agents,” explained Alex Gunnarson, co-founder of Onton. “We’re focused on a different question: what should those agents trust? As commerce moves from clicks to conversations and eventually autonomous actions, the quality of information underneath those decisions becomes critical infrastructure.”
The modern internet was built for human evaluation. Reviews, recommendations, forums, and social proof were designed for people, who could apply intuition, context, and skepticism. As AI agents take over more of the discovery and decision-making process, those human filters disappear. What remains is a system acutely vulnerable to manipulation.
Onton’s model addresses this challenge by evaluating not just what a product is, but whether the information surrounding it can be trusted. By interpreting the attributes and intentions behind a user’s preferences, including visual inputs, the platform is able to treat them as meaningful data, rather than potential noise. This results in AI agents that can make purchasing recommendations grounded in information that’s actually trustworthy.
The company’s launch also includes new research examining how leading AI systems currently evaluate product-related information and where they fail when synthetic content and incentivized recommendations enter the process. The conclusions are pointed: existing systems are not equipped to handle the information environment they are increasingly being asked to navigate.
“We believe the next major internet platform will not be defined solely by who has the best model,” said Zach Hudson, co-founder of Onton. “It will be defined by who can provide the most trustworthy foundation for those models to operate on.”
Onton believes that a trust layer for agentic commerce is a precondition, not a feature. Without it, every AI-powered purchasing decision is built on a foundation that can be gamed. With it, agents can actually do what they promise: help people make better decisions, faster, with information they can rely on.
Ontology 1 is available today for users at Onton.com, and on a case by case basis for partners building on the agentic web who need a trustworthy foundation for product discovery and recommendation.
About Onton
Onton is a search and discovery engine for products. Our mission is to help people make decisions they love, instantly. The average shopping journey takes 79 days, and we’re taking that to less than 1. Since our public launch in late 2023, we’ve grown exponentially to guide millions of users to their perfect product.
Also read:
Freehand Raises $75M to Scale AI Teams Managing Supply Chain Spend for Fortune 500 Companies
ZuriQ raises $25.5 million to scale new groundbreaking 2D quantum architecture
Image source: onton.com
