OpenAI launches shopping research – what does it mean for you?

OpenAI has just introduced shopping research in ChatGPT – a new shopping assistant where customers can describe what they’re looking for in everyday language and instantly receive a curated buying guide with product suggestions.

Instead of jumping between ten browser tabs, ChatGPT does the heavy lifting: reading product pages, comparing options, and helping customers filter down to what really matters.

What is shopping research in practice?

Shopping research is a new mode in ChatGPT where users can type things like:

• “Help me find a quiet vacuum cleaner for a small apartment with a cat.”

• “I need a durable work bag under $200.”

• “Suggestions for a Christmas gift for a ten‑year‑old who loves drawing.”

ChatGPT asks follow‑up questions, searches the web, reads open product pages and reviews, and assembles a buying guide with recommendations, including direct links back to the store where the purchase is completed.

Here’s the important part: shopping research is about helping customers discover the right products and make better decisions, not about replacing your checkout. Customers still click through to your store and complete the purchase as usual. For merchants, shopping research is an additional discovery layer where new customers can find your products, rather than a new channel that requires rebuilding everything.

For Norce customers

If you’re selling through a Norce‑based solution, shopping research means your products can become part of the customer journey inside ChatGPT, provided that ChatGPT can access your webshop and understand what your products actually are.

Step 1: Let ChatGPT into your webshop

For your products to appear, ChatGPT needs to visit your site just like a regular customer, but via a background agent.

• Review whether you currently block “unknown” bots or headless traffic.

• Implement OpenAI’s recommendations to allowlist the ChatGPT agent in WAF/CDN.

• Test by asking real shopping questions in ChatGPT and see if your product pages show up.

Details on allowlisting can be found in OpenAI’s documentation: ChatGPT Agent Allowlisting

Step 2: Make your product pages clear for both humans and AI

Shopping research reads your product detail pages (PDPs) much like a human would. That makes “good PDPs” more critical than ever:

• Clear product names – descriptive titles work better than internal codes.

• Concise summary at the top – one sentence answering “What is this?” and “Who is it for?”.

• Structured information – bullet points and spec tables make comparison easier.

• Honest, helpful text – acknowledging both strengths and trade‑offs builds trust and gives the model more to work with.

If you already use structured data (e.g. schema.org/Product) for SEO, that’s a plus for AI‑driven shopping too. But the most important thing is that the page is logical and easy for a human reader, do that well, and ChatGPT will understand it too. 

What does this mean for the Norce ecosystem?

With Agentic Commerce, we’re entering the next big shift in ecommerce. shopping research in ChatGPT is the first tangible proof that this development is happening now.

Norce Commerce is built for exactly this world: API‑first, headless, and with a strong focus on product data that works for both humans and machines. This gives our merchants and partners a solid foundation to benefit from agentic behaviors, from shopping research today to more advanced agent‑driven buying journeys in the future. 

What you can do right now

• Check in with your Norce partner and make sure ChatGPT traffic isn’t being blocked.

• Review PDPs in a key category – titles, summaries, bullet points, price and stock visibility.

• Plan ahead with your Norce partner on how shopping research and future agentic behaviors fit into your overall ecommerce strategy.

Want to learn more about shopping research?

• OpenAI: shopping research

• ChatGPT Agent Allowlisting