San Francisco-based Onton, an artificial intelligence-powered shopping platform, closed a $7.5 million funding round led by Footwork with participation from Liquid 2, Parable Ventures, and 43, bringing total capital raised to approximately $10 million as the company prepares to expand beyond its furniture origins into apparel and consumer electronics.
The funding announcement comes as Onton reports explosive user growth, scaling from 50,000 monthly active users to over 2 million while processing millions of searches and image generations. The company rebranded from Deft earlier in 2025, citing confusion around the original name and difficulty securing a premium domain.
Addressing E-Commerce Decision Fatigue
According to Onton's research, the average consumer takes 79 days to complete a single purchase decision—a timeline that continues lengthening as internet commerce becomes saturated with SEO-optimized listings, conflicting reviews, and automated content that obscures genuine product information.
"Shopping has quietly become one of the hardest problems on the internet," said Zach Hudson, CEO and co-founder of Onton. "People deserve a way to shop that feels intelligent, transparent, and effortless. Onton is designed to remove the friction that slows everyone down and to give users absolute confidence in their choices."
Co-founder Alex Gunnarson spent 30 hours searching for a mid-century gray couch with wood trim before recognizing the widespread nature of this problem. He connected with Hudson, who had been developing Rcmmd and studying trust in online reviews, at a Y Combinator Startup School event. The pair built early versions, won Pioneer, joined the On Deck Fellowship, and scaled monthly active users to over 1 million with just four employees at the start of 2025.
Neuro-Symbolic Architecture Differentiator
While major technology companies including OpenAI, Google, and Amazon invest heavily in AI shopping assistants, and startups like Perplexity, Daydream, and Cherry build product discovery businesses, Onton differentiates through its core technology: neuro-symbolic architecture.
Hudson explained that while large language models excel at guessing probable intent, they struggle with hallucination problems that undermine e-commerce applications. Onton's approach combines neural networks with explicit rules and ontologies to impose logical consistency—a hybrid method gaining traction in research settings at institutions including IBM Research and Stanford University.
The practical application allows Onton to infer attributes missing from product descriptions and understand synonyms across retailers. When users search for "pet-friendly furniture," the system recognizes that polyester fabrics offer stain and scratch resistance, learning these correlations through accumulated searches rather than requiring explicit product descriptions to contain such information.
"Our tools learn these things through every single search and become smarter at a faster rate," Hudson said.
Visual Discovery Beyond Chat Interfaces
Onton's platform extends beyond traditional chat-based interfaces to include what the company calls an "infinite canvas" for product discovery. Users can upload images of their spaces, generate AI-powered room designs, or describe desired outcomes, with Onton matching them to purchasable products.
The visual approach reflects Hudson's belief that complex shopping decisions require multiple input methods. Users can combine existing images with found products for ideation or upload room photos and request furnishing recommendations.
These approaches yield conversion rates 3 to 5 times higher than traditional e-commerce sites, according to company data. Over 20 percent of users engage weekly, with power users conducting more than 100 searches and generations monthly.
Expansion Strategy and Competitive Landscape
The new capital will fund category expansion, team growth from 10 to 15 employees focusing on engineers and researchers, and continued development of Onton's knowledge graph and data pipeline. The company has already begun building its apparel catalog, with consumer electronics planned subsequently.
User testimonials highlight practical impact. Customers report Onton cutting months from decision cycles, confirming product uniqueness, and enabling confident purchases without endless research loops.
The move into apparel and electronics positions Onton against established AI shopping platforms and well-funded technology giants. Success will depend on whether neuro-symbolic architecture's logical consistency advantages translate across product categories with different attribute complexity and decision dynamics than furniture—where Onton has demonstrated initial product-market fit through its 40-fold user growth.
As AI-powered commerce transitions from experimental features to mainstream shopping infrastructure, Onton's expansion tests whether technical differentiation through reduced hallucinations and improved attribute inference can capture market share against competitors with substantially larger research budgets and existing user bases.