ChatGPT ai Searches Google Shopping to Create its Recommendations 26

The TL;DR: Prioritize Google Shopping for Product Optimization

It was previously thought that the data from ChatGPT's query fan-out (the background Google searches ChatGPT conducts to create a comprehensive response) was the main driver for product inclusion.

However, our experiment shows this is not true. 

ChatGPT generates additional shopping queries sent to Google Shopping. Generally, the results from Google Shopping influence the final products included in ChatGPT's response. 

The usual query fan-out still happens, but it informs the conversational response alongside the product selection. 

Key takeaways:

1. ChatGPT executes two sets of fan-out queries for responses with product carousels. The first set is contextual, used to form the written response. The second set consists of Google Shopping searches, which ChatGPT uses to refine the results.

2. Ecommerce brands should prioritize optimizing for Google Shopping. Products that rank highly there are likely to be included in ChatGPT Shopping recommendations.

How We Conducted Our ChatGPT Shopping Experiment

We aimed to explore the fan-out queries that shape the final answers, first testing whether ChatGPT generates these shopping queries, and then whether the resulting products matched. 

Thus, the first step was to investigate the fan-out queries taking place. 

Step 1: Requesting Product Recommendations from ChatGPT 

We logged into ChatGPT and requested product suggestions with specific criteria. The goal was to interact with the platform as a user would and provide some guidelines for the resulting products. 

For example, we used: “best budget Android phones with great cameras”.

Step 2: Identifying the Fan-Out Queries

You can use a tool like Semrush Enterprise AIO to automatically identify these hidden background searches, but it's also possible with Chrome Dev Tools. Here's how: 

1. Open Chrome Dev Tools 

2. On the Network tab > Fetch/XHR, filter using the conversation URL that ChatGPT creates (the final part of the URL) starting with a number

3. Click the refresh button to reload the conversation and capture the results

4. Use CMD+ F to search the dev tools panel for “search_model_queries” 

5. Here you’ll find the query fan-outs under “queries”

In this case, we have: "best budget Android phones with great cameras 2025" and "what defines a budget Android phone and which budget phones have good cameras 2024 2025".

Step 3: Uncovering the Shopping Fan-Out Queries

Shopping fan-out queries are hidden with an additional layer of encoding.

This process is slightly more complex but can be done as follows:

1. In the same file as before, search for “id_to_token_map”.

2. Locate the text beside this that starts with “ey”. This is a Base64 data snippet that needs decoding.

3. Copy the entire snippet (around 500 characters) without the beginning and ending quote marks.

4. Paste the data into a free tool like Base64 Decode to make it readable and reveal the shopping fan-out queries.

Here's the result:

The key element appears after “query” where it states “cheap+android+smartphones+good+camera+2025”. This is the shopping fan-out query. Typically, only 1-2 unique queries are found here.

Step 4: Comparing the Results

Now equipped with all the information, we can compare results by entering the shopping fan-out query into Google Shopping. Ensure that the locale-location settings match in ChatGPT and Google Shopping for accuracy. 

Here's what we observed for the Android phone example:

Google Shopping

ChatGPT

The first two entries appear in both Google Shopping and ChatGPT. The retailer, title, and price information match exactly. 

The Findings: Shopping Query Fan-Out as ChatGPT’s Additional Search Layer

After conducting the experiment 100 times, we found that the top ChatGPT product appeared in Google Shopping’s top 3 results 75% of the time. There was also significant overlap with the second and third results.

Why Does ChatGPT Utilize Google Shopping Results? 

ChatGPT relies on Google Shopping Results due to their rich data source. This isn’t just a product selection, but a library containing reviews and live pricing info. Products can be recommended confidently with accurate prices and retailers. 

While ChatGPT is working to move away from Google and Google Shopping (evident from the recent Etsy integration and new Shopping Research features), it can't yet match this user experience. Google's ecosystem is vast.

Accurate live pricing info is crucial, especially when prices can change rapidly during ecommerce events like Black Friday.

Implications for Ecommerce Brands

This means ecommerce brands need to ensure their shopping feeds are current, particularly on Google Shopping, as ChatGPT and other platforms strive to develop their own systems.

For product queries in ChatGPT, the logic appears split into: 1. Retrieving buyer-guide context and 2. Retrieving products from Google Shopping.

Currently, optimizing your Google Shopping results is critical for appearing in the ChatGPT product carousel. 

Brands should also prepare for the future. In-chat purchasing and agentic e-commerce will continue to grow. Developments like Google’s Universal Commerce Protocol (UCP) will streamline these processes. 

We're moving towards a world where transactions occur directly within the chat, without visiting your website.

About the Experiment

This experiment analyzed the fan-out queries of 100 prompts. Each prompt was run 5 times, recording the most common top products in the carousel to account for the probabilistic nature of LLMs. 

Although this is a small sample size, it provides a starting point for AI shopping optimization. We’ll continue to explore shopping query fan-out with larger datasets for further insights.

You can conduct this test yourself. While we used a logged-in account, results may vary depending on whether you're logged in, have a free account, or a paid account, as seen in similar ChatGPT and Google experiments we’ve conducted.

nano banana pdf 2026 how to fast learn explained 1,26,1

Buy Now

10 prompts best for nano banana winter 26 pdf

Buy Now