With the holiday shopping season approaching, OpenAI and Perplexity have both rolled out new AI-powered shopping features, embedding product discovery directly into their chatbots. The launches signal a major push by the two AI giants into consumer commerce — a move that could reshape how shoppers research and compare products online.
The capabilities announced by the companies look strikingly similar. OpenAI says users can now ask ChatGPT to help them identify very specific products, such as “a new laptop suitable for gaming under $1000 with a screen that’s over 15 inches.” Users can even upload photos of premium fashion items and request more affordable alternatives. Perplexity, meanwhile, is leaning on its long-term memory system, highlighting that its chatbot can personalize recommendations based on details it already knows — like a user’s job or location — adding a contextual layer to product searches.
This rapid shift comes as Adobe forecasts that AI-assisted online shopping will surge 520% this holiday season, a trend that could benefit specialized AI commerce startups like Phia, Cherry, and Deft (now Onton). But with OpenAI and Perplexity aggressively expanding into the same territory, the central question emerges: can these niche players survive the competition?
Zach Hudson, CEO of AI interior design platform Onton, believes they can — as long as specialization remains their edge. “Any model or knowledge graph is only as good as its data sources,” he told TechCrunch, arguing that general-purpose systems like ChatGPT and Perplexity rely heavily on mainstream search indexes. As a result, he says, “they’re really only as good as the first few results that come back from those indexes.”
Daydream CEO Julie Bornstein echoes this view. She notes that fashion search, in particular, has long fallen short because it lacks deeper contextual intelligence. “Fashion … is uniquely nuanced and emotional — finding a dress you love is not the same as finding a television,” she said. Effective fashion discovery, she adds, requires domain-specific data and merchandising logic that captures silhouettes, fabrics, and evolving personal style — something general models cannot easily replicate.
Startups like Onton build their own customized datasets, cataloging hundreds of thousands of products with structured attributes to train more accurate models. But Hudson warns that without this level of depth, “it’s very hard to see how a startup can compete with the larger companies.”
OpenAI and Perplexity, however, have one major advantage: scale. Their existing user bases allow seamless integration of shopping tools, and partnerships with major platforms — Shopify for OpenAI and PayPal for Perplexity — enable in-chat checkout, bypassing the affiliate-based models many startups rely on.
This shift also hints at a larger strategy: as AI companies search for sustainable revenue models, e-commerce — and potentially advertising — may become central. But Bornstein cautions that this could worsen existing search issues, emphasizing that “vertical models — whether in fashion, travel, or home goods — will outperform because they’re tuned to real consumer decision-making.”
