AI and product discovery: Q1 2025 data

How AI assistants evolved (Q1 2025)

AI-based assistants are continuing to advance, leveraging sophisticated machine learning algorithms and vast datasets to provide more accurate and contextual responses. A key evolution is their deeper integration with real-time web search capabilities and direct shopping options or purchase links.

ChatGPT, for example, has further refined its ability to browse the web and suggest relevant products, and through the partnership between OpenAI and Shopify, it is about to integrate direct checkout within the chat. Claude introduced web search functionality on March 20, 2025, broadening its product discovery use cases. Perplexity maintains its integration with Shopify, simplifying the path from information to purchase. Gemini is progressively adding product search features, although its strategy still centers around Google SERP’s AI Overviews. Meanwhile, Mistral and DeepSeek are showing early-stage developments in product discovery capabilities, though current traffic volumes remain low.

Ecommerce impact (Q1 2025)

The latest aggregated ecommerce traffic analysis reveals a sharp increase in referral sessions coming from AI tools (ChatGPT, Perplexity, Gemini, Claude, Mistral, and DeepSeek)—up 721.3% in Q1 2025 compared to Q4 2024.

This traffic surge is also accompanied by positive signals in user engagement and buying intent. The average session duration increased by 12.9%, indicating greater engagement from users arriving via AI. There was also a rise in the add-to-cart (ATC) rate, which now stands at 3.3%, marking a 101.2% growth. It’s worth noting that this increase may have been influenced by the winter sales in Q1 2025, which typically encourage users to add products to their cart.

Top AI referral sources (Q1 2025)

Analyzing data from the first quarter of 2025, ChatGPT remains the leading generator of AI-driven traffic, accounting for 89.3% of sessions and 93.75% of revenue.

Following at lower volumes are:

• Perplexity: generating 6% of sessions and 4.16% of revenue.
• Gemini: contributing 2.2% of sessions and 2.09% of revenue.

Lower traffic volumes were also observed from Mistral, Claude (post-March launch), and DeepSeek, suggesting potential future diversification in AI traffic sources. Copilot still shows limited ecommerce referral data.

Industries most affected by AI traffic (Q1 2025)

The industries receiving the highest share of AI-driven traffic include:

• Hard Goods (Consumer Electronics): 34,60% of sessions
• Fashion: 26,64% of sessions
• Sports: 15,16% of sessions
• Home & Design: 8,58% of sessions
• Luxury: 6,48% of sessions
• Retail & Shops: 5,43% of sessions
• Beauty: 2,96% of sessions
• Kids: 0,16% of sessions

These insights confirm and expand on previously observed trends, showing strong interest in consumer electronics and fashion.

What the Q1 2025 data tells us

Data from the first quarter of 2025 outlines a more consolidated phase in the adoption of AI for online product discovery.
The increase in average session duration points to higher user engagement, while the growth in the add-to-cart rate (excluding the impact of seasonal sales) suggests a shift from an exploratory phase to a more purchase-oriented usage.
In the coming months, we will continue to closely monitor the evolution of various AI platforms, user adoption, and their impact across the entire conversion funnel — also analyzing the true influence of Google’s AI Overview.