OpenAI and Perplexity are launching AI shopping assistants, but competing startups aren’t sweating it (techcrunch.com)

OpenAI and Perplexity are rolling out AI-powered shopping assistants that let users ask conversational queries (e.g., “find a gaming laptop under $1,000 with a 15+ inch screen”) and even use photos to request lookalike items at lower price points. Perplexity highlights its memory feature to tailor recommendations to known user context, while OpenAI and Perplexity have tied into commerce infrastructure (Shopify and PayPal integrations) so users can complete checkout inside the chat. Adobe projects AI-assisted online shopping could surge ~520% this holiday season, making e-commerce a prime monetization route for compute-heavy LLM platforms — through partnerships, in-chat checkout, or retailer-paid placements. But competing startups aren’t necessarily boxed out. Founders of niche shopping AIs argue that vertical specialists win on data quality and merchandising logic: domain-tuned models trained on proprietary, curated catalogs (fashion silhouettes, furniture specs, interior design pipelines) better capture nuance than general LLMs that largely repackage search-index results. That creates a technical divide — broad LLMs offer convenience and scale, while vertical models offer higher precision and conversion because of cleaner datasets and task-specific training. The near-term landscape will likely be hybrid: big LLMs drive reach and checkout experience, while vertical startups differentiate on domain data, UX and recommendation accuracy.