OpenAI has just launched a new feature in ChatGPT called Shopping – and for my point of view, it could have significant consequences for both search behavior and online advertising.
In short: you can now use ChatGPT to search for products, compare them, and get recommendations – similar to how you’d use Google. But here, a language model guides the experience.
However, this is more than just another feature. It’s a direct challenge to one of Google’s most profitable business areas: shopping-related searches.
Google’s cash cow: ads and e-commerce
To understand the potential of ChatGPT’s new direction, we need to look at the numbers:
- In 2023, Google generated over $237 billion USD in ad revenue (Statista).
- Most of this came from Google Search and Shopping Ads.
- “Commercial intent” searches (users looking to buy something) are the most valuable clicks in Google’s ecosystem.
This is exactly the segment ChatGPT Shopping is targeting. Not necessarily by replacing it – but by offering an alternative buying journey, where the AI acts as an advisor instead of just showing 10 blue links and four ads.
How ChatGPT Shopping works
When you ask a question like “best running shoes for wide feet,” ChatGPT now returns a set of curated product recommendations – with images, descriptions, reviews, and prices.
This data is pulled from the new Shopping Graph, which combines partner feeds (e.g. Instacart and Klarna) with real-time data from open sources.
The feature isn’t just a list. It uses natural language understanding to evaluate which products best match your intent – and filters information based on preferences, needs, and context. It’s a kind of personal product assistant that goes far deeper than traditional keyword matching (Wired).
What does it mean for e-commerce professionals?
If you sell products online – whether you’re a brand, DTC shop, or agency – you need to be aware that product search is shifting. Not away from Google entirely (yet), but into an alternative ecosystem where visibility may not come from paid ads.
That’s why it becomes critical to ensure:
- Structured product data: name, price, images, descriptions, review scores, variant details
- Schema markup: make sure your shop provides correctly structured data that LLMs can read
- Product descriptions with semantic depth: not just keywords, but meaningful content that can be interpreted in context
Is this a real threat to Google?
In short: Google still has the first-mover advantage, the ad platform, and the world’s largest search database. But user expectations are changing – especially among younger audiences.
Google dominates because it owns both traffic and transactions. But if ChatGPT becomes a top-of-funnel search assistant, and users start their research there, Google loses influence over the buying journey.
That’s exactly the shift that could eventually impact their ad revenue – and their core business.
What’s the current status?
ChatGPT Shopping has been rolled out to Plus users (GPT-4) and currently works mainly in the English version. It’s being improved continuously, and new partners are being integrated. It’s still in beta – but a very ambitious one.
OpenAI isn’t trying to “replace” Google, but rather become the first contact point for users who want an answer – not just a list.
What should you do now?
- Check whether your products are visible via feeds, schema, and proper indexing
- Rethink your product descriptions – are you writing for algorithms, humans, or LLMs?
- Keep an eye on ChatGPT’s evolution in the GPT Store, browser plugin, and native shopping features
This isn’t about choosing ChatGPT or Google. It’s about understanding that users are searching differently – and future visibility depends on being found by a language model, too.
Conclusion
With the Shopping feature, ChatGPT is moving seriously into commercial search – and that brings new demands for those working in marketing, SEO, and e-commerce. If you want to be part of where the buying journey begins tomorrow, now is the time to prepare.
AI and search are merging – and the question isn’t whether your products are shown. The question is whether they are chosen by the model.
Comments