Online shopping was supposed to make people’s lives easier, providing instant access to any product they might need or want. The reality, most people appear to feel, is currently very different, because searching for the items they covet is often so frustrating. Research from Nosto highlights a string of problems, from search processes that generate endless irrelevant results – often paid for by merchants – to products that are displayed despite being out of stock.

Consumers aren’t the only ones who are fed up. Retailers are also finding the going tough: one Google study has put the global cost of search abandonment – when a shopper looks for a product on a retailer’s website but can’t find what they want – at more than $2 trillion a year.

San Francisco-based start-up Onton believes it can solve this problem. Founded in 2023 as a rival search and discovery engine for e-commerce shoppers, the company is today launching Ontology 1, a new model that will underpin artificial intelligence-powered search.

The average shopper currently takes weeks to complete an online shopping journey, says Onton co-founder Zach Hudson. “That’s the problem we set up to solve – can we bring the decision-making time of online shopping down to nearly zero,” he says. “By doing so, our vision is to be the starting point of every shopping journey.”

It’s an ambitious goal, given that Onton has, to date, focused on just one niche of e-commerce, home furnishings. Still, Hudson points out that Onton’s own engine is on target to complete 200 million searches on behalf of consumers before the end of the year – five times’ more than in 2025 – and that it will shortly launch in new categories, including apparel.

The company hopes Ontology 1 will help it build quickly on these foundations. The goal is for consumers to be able to search in a much more human way – through natural language queries, for example, and by using pictures of items they want something similar to, or to get a match with. The model is also intended to span both product discovery and decision-making, providing shoppers with suggestions and recommendations that reflect their individual preferences and needs much more closely, as well as trusted sources of information found elsewhere online.

Onton is hedging its bets somewhat. The difficulty of securing brand recognition in an e-commerce market dominated by giants such as Amazon and Google Shopping has seen it add an enterprise offering to its value proposition. Retailers and e-commerce merchants will be able to link to Onton’s new model, white labelling the service to power their own search engines for shoppers.

“Today, you have to fullfill the shopping journey for wherever the user is within it because e-commerce search hasn’t really changed in 20 years,” adds Hudson. “It’s been a single-threaded search that was built for our parents' generation of online shopping; the reason why it's been stuck that way is we didn't have the technology to do it.”

Generative AI is the missing link, Hudson suggests, but argues that non-specialist large language models (LLMs) can’t deal with the messy data of the ecommerce sector. “We didn't set out to build this model,” he adds. “We built a research team to explore what would work for our use case, and what would give users a better shopping journey, and we stumbled on something incredible.”

Maybe, but Onton will have to fight off a wide array of competitors to live up to Hudson’s vision – a tough ask for an early-stage business that has so far raised only $10 million or so of capital. Tech giants investigating AI-enhanced search include everyone from Google to Shopify, plus all the big players in LLMs. Smaller players range from Fast Simon to Rezolve.

Still, there is clearly demand for what these organizations have to offer. “AI-powered search engines are becoming more and more popular in both consumer and business settings due to the trend of users looking for quick responses in a conversational and contextually appropriate way,” concludes a report published earlier this year by market research group SNS Insider.

The AI search market is set to grow at a rate of 17% a year over the next decade, SNS estimated, with ecommerce driving that acceleration.