AI Shopping Heats Up as OpenAI and Perplexity Crash the Party. The Specialists Say: Let Them Try
Two general-purpose chatbots want to reinvent retail. The niche players think they have the advantage.
OpenAI and Perplexity have stepped directly into the shopping arena, launching AI assistants that help users research products, compare options and receive personalised recommendations inside a chat interface.
The tools operate as conversational concierges, capable of identifying gaming laptops under a certain price, finding low-cost alternatives to designer garments or suggesting items based on the user’s location, job or habits.
It is a confident move into a space that has been steadily filling with startups built around similar ideas.
Despite the noise, specialist AI shopping companies are not alarmed. They argue that the value of a shopping assistant depends on the depth and structure of its data.
General-purpose chatbots lean heavily on broad search indexes, which can be powerful but tend to flatten nuance.
For categories like fashion, furniture or interior design, that gap matters. The details are not simply metadata. They are the entire decision process.
Startups such as Onton and Phia have spent years building domain-specific catalogues to avoid this flattening effect.
They ingest products with clean, structured data and organise them in ways that allow their models to reason about style, texture, context or physical space.
This approach is labour-intensive, but it gives vertical tools a level of precision that generic models often struggle to match.
Fashion experts see the same divide. They describe fashion search as historically weak because it does not handle context or emotion well. Decisions about silhouettes, fabrics, occasions and personal style require structured knowledge that broad models lack.
This is why vertical AI shopping companies believe they can hold their ground. Their systems are trained to understand the real decision-making logic behind purchases.
OpenAI and Perplexity still enjoy major advantages. They already reach millions of users and have secured retail partnerships that allow direct checkout inside the chat interface. This creates a smooth end-to-end experience that smaller players cannot yet match.
The platforms also need sustainable revenue streams to fund their compute demands, and commerce is an obvious route. Retailers may eventually pay to surface products inside AI-generated shopping results.
Even so, vertical AI teams believe the long-term winners will be the tools that understand categories deeply rather than broadly. In their view, precision beats scale, and specialist models remain better suited to the complexity of real consumer choices.
