Big tech companies aren’t merely using AI to help you create or summarize content — they also want you to use it for shopping. OpenAI, Google, and Amazon have bet big on AI assistants that do research for you into new product categories and tell you which ones to buy.
Startups like Perplexity, Daydream, and Cherry have also emerged to build businesses around AI for product discovery. All of these efforts have led customers to do more AI shopping. Onton (formerly Deft), an AI-enabled furniture shopping platform, reports that its user base has grown from 50,000 monthly active users to over 2 million, with millions of searches and images generated.
With that growth as its wind in the sails, today the startup is also disclosing $7.5 million in funding, a Series A led by Footwork, with Liquid 2, Parable Ventures, and 43North also participating. This round brings the startup’s total funding to around $10 million.
With this funding, the company aims to expand into additional categories, such as apparel and, ultimately, consumer electronics.
Earlier this year, the company rebranded from Deft to Onton because people were confused about what the original name meant, and finding a premium domain name was difficult.
Although large language models (LLMs) are good at guessing likely intent, “they haven’t solved a whole lot of stuff in e-commerce,” said Zach Hudson, co-founder of Onton. The founder added that the startup has seen a spike in the time it takes the average consumer to make up their mind about purchasing anything.
For its main technology, the company taps into neuro-symbolic architecture. With this approach, the company can overcome LLM hallucination issues and deliver more accurate, logical search results, Hudson said. The startup’s model, he added, can also learn information from the real world that might not be detailed in a product description.
“Let’s say you are searching for pet-friendly furniture. Our tools understand that if an item contains polyester, it would be more stain- and scratch-resistant, making it more pet-friendly. Our tools learn this stuff through every single search, and they are getting smarter exponentially faster,” Hudson added.
When you’re trying to find something that may be called one thing on this site and another thing on another site, he added, you don’t get good results. The company’s AI model accounts for those scenarios when delivering results.
Onton has added additional input methods and features to support individuals in both short- and long-term decision-making. Now, you can also upload an image or enter a prompt to create what you’re looking for in your house or office setup, and OntoN will help source the furniture to match.
Onton also provides an endless canvas for image generation, allowing you to add existing images and the items you’ve uncovered for inspiration. You can also upload photos of your room and ask the tool to fill it with furniture.
The company believes that, rather than restricting users to chat only, features like these will offer them more ways to get to what they want, even if they don’t know how to describe it perfectly.
The startup claims that, through these methods, it’s managed to convert customers at a rate 3-5x higher than traditional e-commerce sites (which have similar audiences but no trust in the data underlying them).
Because of the technological and interface changes Hudson has made, launching apparel will be easier. The company is also bolstering its supply for the category and will launch the vertical soon. It will be in the company of others in this category, such as Daydream, Aesthetic, and Style.Ai.
The company, which had 3 full-time employees in 2023 and now has 10, is looking to hire engineers and researchers to grow the team to 15.
