Press Release: Where ChatGPT Already Works in Online Shopping – and Where It Does Not

Why It Matters

The findings temper expectations that ChatGPT will replace established digital marketing, highlighting where generative AI can actually boost conversion and where retailers should continue to rely on proven channels.

Key Takeaways

• ChatGPT drives <0.2% of e‑commerce traffic in study.
• Conversion rate exceeds paid social but trails all other channels.
• Higher value seen in information‑intensive, complex‑purchase categories.
• Simple‑product retailers gain limited short‑term benefit from ChatGPT.
• LLM firms adding instant checkout and agentic shopping features.

Pulse Analysis

The hype around ChatGPT as a "Google killer" for online shopping has been tempered by a new empirical study from the Frankfurt School of Finance & Management. Analyzing organic referrals from large language models to 973 e‑commerce sites—collectively generating more than $20 billion in revenue—the researchers discovered that AI‑driven traffic represents under 0.2% of total visits. While the conversion rate outperforms paid social, it still falls short of organic search, email and affiliate channels, underscoring that generative AI is not yet a wholesale substitute for established digital marketing tactics.

Where ChatGPT does add measurable value is in complex, information‑intensive purchase decisions. In categories such as electronics, home appliances, or high‑ticket services, the model helps shoppers structure product information, compare alternatives, and navigate nuanced specifications. This guidance translates into higher conversion efficiency relative to paid social, suggesting that AI can act as a complementary advisory layer rather than a primary acquisition driver. Conversely, for low‑involvement, simple‑product purchases—think everyday consumables—the AI’s influence remains marginal, offering little incremental revenue for retailers.

For businesses, the study signals a strategic pivot: integrate ChatGPT where it enhances the decision‑making journey, but continue to invest in proven channels for volume traffic. LLM providers are already responding, rolling out instant checkout and agentic shopping features to close the purchase‑completion gap. As these capabilities mature, we can expect the AI’s share of e‑commerce traffic to rise, but its role will likely remain that of a specialized assistant rather than a universal traffic source.