OpenAI and Perplexity Target Shopping. Startups Unfazed

OpenAI and Perplexity launched new AI-powered shopping assistant features within their chatbots this week to capture a booming holiday e-commerce market predicted to grow by 520%.

These new tools aim to streamline product research for consumers. For instance, OpenAI allows ChatGPT users to search for highly specific items, such as a gaming laptop under $1,000 with a screen larger than 15 inches, or upload photos of premium fashion items to find budget-friendly alternatives. Meanwhile, Perplexity leverages its chatbot’s memory to personalize recommendations based on user demographics like occupation and location.

Why Niche Startups Aren’t Worried About Tech Giants

Despite the rapid expansion of these tech giants, specialized e-commerce startups remain confident in their competitive edge. Adobe’s projection of a massive 520% surge in AI-driven holiday shopping signals a massive opportunity for startups like Phia, Cherry, and Onton (formerly Deft). Industry leaders argue that general-purpose AI models cannot match the depth of specialized systems.

Zach Hudson, CEO of interior design shopping assistant Onton, highlights the limitations of general Large Language Models (LLMs). According to Hudson, tools like ChatGPT and Perplexity rely heavily on standard search indexes like Google or Bing, restricting their output quality to the top search results. While Perplexity maintains its own search index, Hudson insists that a model’s output is only as strong as its underlying data pipeline.

The Power of Domain-Specific Data

The fashion sector presents a unique challenge that general AI struggle to navigate. Julie Bornstein, CEO of Daydream and a veteran e-commerce executive, points out that search has historically been neglected in the fashion industry. She emphasizes that buying clothing is a deeply emotional and nuanced process, vastly different from purchasing electronics.

Bornstein notes that true fashion understanding requires domain-specific data and merchandising logic. This allows AI to comprehend silhouettes, fabrics, occasions, and how consumers build outfits over time. To achieve this, specialized startups build proprietary datasets rather than attempting to catalog all human knowledge. For example, Onton created a custom data pipeline to index hundreds of thousands of home decor products, ensuring cleaner data for training its models.

The Platform Advantage: Distribution and Checkout

Startups without specialized data pipelines risk being left behind if they rely solely on off-the-shelf LLMs. However, OpenAI and Perplexity possess a major advantage: a massive, pre-existing user base and established retail partnerships. While niche platforms like Daydream and Phia redirect users to external websites to complete purchases, the tech giants offer native checkout experiences through integrations with Shopify (OpenAI) and PayPal (Perplexity).

As these heavily funded AI companies seek profitability, e-commerce offers a lucrative monetization path. Following the playbooks of Google and Amazon, they could eventually allow retailers to pay for sponsored product placement within search results. However, experts warn that prioritizing ad revenue over user experience could recreate the very search issues consumers currently face, giving vertical models in travel, fashion, and home goods a long-term advantage.