Such is the pace of development and disruption in the AI space that it can be tricky to keep up with the milestones as they come and go. One such threshold was crossed recently, with two of the biggest players, OpenAI and Perplexity, both introducing agentic shopping assistants designed to optimise how users browse and buy online. 

These rival assistants work in much the same way, leveraging a natural conversational back-and-forth which allows them to feel out user requirements, refine their recommendations and remember buying histories to find provide a tailored experience. This is all in marked contrast to the lists of non-curated links yielded by traditional search.

The mass rollout of this kind of conversational, agentic AI undoubtedly comprises the most significant evolutionary step in e-commerce since the emergence of smartphones put digital shopfronts in everybody’s pockets. And while the use of autonomous shopping agents is still in the early phase of the adoption curve, recent market analysis by Adobe predicts that the overall use of AI for activities like shopping research and deal finding will see a mammoth 520% year-on-year surge this holiday season. 

The seismic ramifications of the new era are already being felt, with online retailers now compelled to move beyond traditional SEO practices to make themselves visible to AI agents (discussed more in a previous article on the AI takeover of search). We’re also seeing the flourishing of a fresh ecosystem of companies which are developing AI shopping agents tailored to specific retail verticals. This begs the question: can these plucky startups survive – in fact, do they even have a reason to exist – when the likes of ChatGPT and Perplexity now include agents in their interfaces? 

Investors think so, for two strong reasons. First, vertical-focused startups can more easily cultivate high-quality data. As Hampleton Partners’ managing director Jonathan Simnett emphasised in a recent interview on the critical significance of data management to the AI revolution, “Open-source LLMs are rapidly commoditising, and the next differentiator in AI will not be who has the best model, but who has the cleanest, most traceable data pipelines.”

By developing agents which are trained on smaller, niche-specific, proprietary datasets, startups can maintain an edge over general-purpose models, providing more relevant results to users. Indeed, high quality, carefully curated data promises to be the key differentiator and the lifeblood of ambitious startups. The CEO of interior design shopping platform Onton emphasised this point in a recent interview, warning that “If you’re using only off-the-shelf LLMs and a conversational interface, it’s very hard to see how a startup can compete with the larger companies.”

Data and other backend considerations aside, the user experience of more specialised shopping platforms is also likely to be critical to their success. For example, a fashion platform which presents seasonal edits, analysis of silhouettes and materials, hyper-personalised advice and loyalty programmes, all within a distinctively styled app or website augmented by active social media community building, will provide a more immersive and emotionally intelligent experience than a shopping assistant churning out suggestions in a neutral, general-purpose interface. 

This of course isn’t to say the likes of ChatGPT, Google and Perplexity won’t enjoy the lion’s share of agent interactions. But, given the sheer size of the agentic e-commerce market (which in the US alone could account for up to USD 386bn in e-commerce spending by 2030), there’s room for a diverse range of assistants to co-exist. Well-designed, vertical-specific platforms may well thrive as the boutique options for discerning shoppers who value more distinct and nuanced retail experiences. Here are some which investors have been paying attention to.

One of the most talked-about companies in this space is Daydream, which launched in June having raised USD 50m in funding. A fashion-focused platform, it exemplifies how vertical-focused startups can differentiate through deep domain expertise (it has partnerships with over 8,000 fashion brands) and an eye-catching aesthetic. Daydream has a suitably glossy, glamorous feel, with its CEO stressing the importance of a “nuanced and emotional” user experience for consumers looking to buy lifestyle items like haute couture. 

A furniture and home décor shopping platform, Onton has seen its user pool grow from 50,000 to more than two million per month – an impressive leap which will be further fuelled by its recent USD 7.5m fundraise. As well as providing smart suggestions through its conversational interface, the platform allows users to upload pictures of their properties to digitally populate them with potential purchases. The agent’s recommendations derive from a proprietary data pipeline designed to slash the time it takes to find the right items and boost conversion rates, which according to the company are three to five times higher than the industry benchmark.