Onton’s Ontology 1 shows why ordinary keyword and vector search fails when shoppers ask for nuanced, multimodal needs like pet‑friendly furniture that fits a room or lighting that creates a specific mood. Traditional engines rely on surface tags and learned embeddings, which miss hidden attributes such as material durability, weave tightness, or construction quality. They also treat seller‑provided labels as truth, even when those labels are missing or misleading. As a result, precision drops on long, requirement‑heavy queries and on image‑only searches that major platforms do not support.
Ontology 1 builds an explicit, inspectable world model instead of burying patterns in weights. When a concept like “pet‑friendly” has no direct entry, the system reasons from observable properties—fiber type, weave, cleanability—and flags contradictions in the data. It weighs source reliability, discounts gamed listings, and reuses learned reasoning for future queries such as “pet‑friendly chair” or “cleanable blue couch.” The model runs on a custom graph database that delivers far higher throughput per core than conventional sparse‑matrix tools, and a GPU version scales even further.
Deployability is partnership‑based: Ontology 1 is live at Onton.com for end users, with case‑by‑case access for retailers building agentic commerce experiences. There is no public downloadable weight set, API, or pricing sheet; adoption works through direct collaboration rather than a pip install. Benchmarks on the Subtext‑Decor‑90 set show a mean precision@10 of 0.630, outperforming Google Shopping (0.543) and Amazon (0.469) while indexing roughly 1% of their catalogs. The approach improves relevance for home decor and furniture today, and the method is said to generalize to other verticals and non‑product data.
For retailers losing relevance on complex, conversational searches, Ontology 1 offers a practical path to better results without rebuilding entire stacks—just a partnership that adds a neurosymbolic reasoning layer on top of existing catalogs.
#AI #Product #Search #Ecommerce #Innovation #Tech