OpenAI and Perplexity are launching AI shopping assistants, but competing startups aren’t sweating it

With holiday shopping on the horizon, OpenAI And Bewilderment both announced AI shopping features this week, which can be integrated into their existing chatbots to help users research potential purchases.

The tools are clearly similar. OpenAI suggests that users can ask ChatGPT for help finding a “new laptop suitable for gaming under $1,000 with a screen larger than 15 inches,” or they can share photos of a high-end piece of clothing and ask for something similar at a lower price.

Baffling, meanwhile, is how the chatbot’s memory can augment shopping-related searches for its users, suggesting that someone could ask for recommendations tailored to what the chatbot already knows about them, such as where they live or what they do for work.

Adobe predicted that AI-enabled online shopping will grow 520% ​​this holiday season, which could be a boon for AI shopping startups like Phia, Cherry or Deft (rebranded Onton) – but are these startups at risk as OpenAI and Perplexity delve further into AI shopping experiences?

Zach Hudson, CEO of the interior design tool Ontonthinks AI shopping startups with a specialized niche will still provide users with a better experience than generic tools like ChatGPT and Perplexity.

“Any model or knowledge graph is only as good as its data sources,” Hudson told TechCrunch. “Right now, ChatGPT and LLM-based tools like Perplexity leverage existing search indexes like Bing or Google. That really only makes them as good as the first few results that come back from those indexes.” (Bewilderment noted to TechCrunch that it has its own search index.)

Daydream CEO and former e-commerce executive Julie Bornstein agrees. She noted to TechCrunch last summer that she always considered search “the forgotten child” of the fashion industry because it has never worked particularly well.

“Fashion… is uniquely nuanced and emotional – finding a dress you love is not the same as finding a television,” Bornstein told TechCrunch on Tuesday. “That level of understanding for fashion shopping comes from domain-specific data and merchandising logic that understands silhouettes, fabrics, occasions and how people build outfits over time.”

AI retail startups are developing their own data sets, so their tools are trained on higher quality data – something that’s easier to achieve if you’re trying to catalog fashion or furniture than the sum of all human knowledge.

In Hudson’s case, Onton developed a data pipeline to catalog hundreds of thousands of interior products in a cleaner way, allowing the internal models to be trained with better data. But if AI retail startups don’t pursue that level of specialization, Hudson thinks they will be overshadowed.

“If you’re just using off-the-shelf LLMs and a conversational interface, it’s very difficult to see how a startup can compete with the larger companies,” Hudson said.

The advantage for OpenAI and Perplexity, however, is that their customers are already using their tools – and their large presence allows them to make deals with major retailers from the start. While Daydream and Phia direct customers to retailers’ websites to complete their purchases — sometimes earning affiliate revenue — OpenAI and Perplexity have partnered with Shopify and PayPal, respectively, allowing users to check out within the conversational interface.

These companies, which rely on massive amounts of expensive computing power to operate, are still trying to find a path to profitability. If they are inspired by Google and Amazon, it makes sense to consider e-commerce as an option; retailers could pay them to advertise their products in search results.

But ultimately, that could only exacerbate existing problems customers have with search.

“Vertical models – whether for fashion, travel or home goods – will perform better because they are aligned with real consumer decision-making,” Bornstein said.

Additional reporting by Ivan Mehta. Updated 11/26/25, 11:30 AM ET with commentary from Perplexity.