AI Shopping Comparison: Perplexity Leads Listings, GPT Wins Specs

With artificial intelligence (AI) becoming more integrated into daily life, many consumers are using AI chatbots to shop online. To see how effective they really are, we tested three popular AI toolsโ€”ChatGPT, Perplexity, and Geminiโ€”across three use cases: buying a laptop, selecting sunscreen, and finding a Fatherโ€™s Day gift.

We input the same prompts into all three platforms and compared the results.

Reporter Cho testing AI
Photo by Naki Park

Perplexity: Best for visual listings and ease of use

Perplexity offered the most user-friendly experience. It organized results visually, listing products with photos, features, and summaries. For example, it described an MSI laptop as a budget-friendly option and highlighted the HP model for its vivid display. Sunscreen products came with notes on hydration, light texture, and no white cast.

It also included user reviews and star ratings from major retailers like Walmart and Best Buy, and premium users could click โ€œbuyโ€ and check out instantly using saved cards. However, Perplexity fell short in real-time price comparison. For instance, a featured sunscreenโ€”Beauty of Joseonโ€”was listed at over $15, while it was available elsewhere for about $14 with coupons when buying a two-pack.

ChatGPT: Best for detailed product specs

ChatGPT excelled at technical comparisons. It laid out laptop specsโ€”like CPU modelsโ€”in detail and listed sunscreen ingredients for comparison. Although it cannot perform live price tracking, it linked to price comparison websites and offered general purchasing advice. When pushed for prices, it responded with disclaimers like โ€œnot real-time data.โ€

For Fatherโ€™s Day gifts, ChatGPT stood out by suggesting creative options and even including examples of card messages.

Gemini: Least helpful in current form

Gemini, Google’s AI assistant, lagged behind in usability. Its product comparisons were vague and didnโ€™t clarify pros or cons. It offered minimal links and gave generalized answers like โ€œprices vary by configuration.โ€ Even when prompted with specific product names, it provided basic links without summaries.

For Father’s Day gifts, Gemini recommended generic items like supplements, without much personalization.

Prompt quality directly impacts results

One clear takeaway: AI responses improve dramatically with more specific prompts. When we provided details like the recipientโ€™s age, hobbies, or location, all tools delivered more tailored suggestions. A well-crafted prompt is key to getting useful shopping advice from any AI platform.

BY WONHEE CHO [cho.wonhee@koreadaily.com]