Just in time for the holiday shopping season, OpenAI has launched a new tool that it anticipates will help you find exactly what you want.  

Key takeaways. 

• OpenAI has introduced ChatGPT Shopping Research within its platform.  
• The functionality helps users discover products and deals.  
• The OpenAI shopping tool is available on both mobile and web platforms for all logged-in ChatGPT users and for Black Friday AI deals.  

The holiday season has arrived, and the most significant shopping event of the year, Black Friday, is only a few days away. Considering this, OpenAI has introduced a new feature in ChatGPT to assist with your business shopping needs at no extra cost.  

Although the company has previously released AI Agents that focus on automating the order-placing process, such as OpenAI Atlas and Instant Checkout, this ChatGPT shopping research feature is designed to help you discover what you want to buy first, acting as an expert personal shopper.  

In a manner akin to deep research, when provided with a product description, ChatGPT will now navigate the internet to compile a guide for you. Additionally, it will pose a series of clarifying questions, drawing on context from previous conversations and product reviews to formulate your guide.  

What is the purpose of ChatGPT shopping research (how it operates)? 

Shopping Research is intended to serve as an assistant that can generate a customized shopping experience tailored to your individual criteria and requirements in just a few minutes. OpenAI has stated that the outputs of this research can assist with a range of tasks, including locating a product that fulfils specific criteria, such as: “Assist me in finding a smartphone with over 18 hours of battery life priced at under $1,500.”  

Additional examples encompass:  

• Identifying duplicates or look-alikes of a product  
• Comparing various products  
• Comprehensive trade-off list that is customized to your particular needs  
• Discovering product deals  
• Aiding you in selecting gifts for individuals on your list  

Furthermore, Amazon is not keen on Perplexity because it uses AI to conduct browser shopping on your behalf.  

The entire process is driven by a variant of ChatGPT-5 Mini, specifically trained by OpenAI for shopping-related tasks. The organization has indicated that it was trained to analyze trusted websites, reference credible sources, integrate information from multiple sources, and refine its prompts in real time.  

Compared with other ChatGPT models, such as ChatGPT-5, ChatGPT Thinking, or ChatGPT Search, the AI product finder excels at product accuracy. However, OpenAI has recognized that it sometimes errs regarding product specifics, including prices and availability, and advised users to always verify its findings.  

How to use ChatGPT for shopping research (and access it) 

I found the AI product finder to be both interactive and intuitive. To begin with, all ChatGPT users who are logged in, including those with free, Go, Plus, and Pro subscriptions, can either post a shopping query to automatically activate the feature or choose the shopping research option from the toolbox menu.  

Your initial prompt specifies what you would like it to accomplish. Subsequently, ChatGPT will ask follow-up questions relating to your search, such as your budget or the features that matter most to you. It will also leverage the context it has about you, provided you have enabled those personalization settings to customize the responses accordingly.  

As it performs the research, it will showcase sample products it has identified. For each product, you can indicate whether you are interested and explain your reasoning, which will further guide the research. This aspect of the feature was my favorite, as it felt like a fun Tinder-like experience, letting you swipe through to indicate your preferences.  

After a few minutes, you will be presented with a personalized buyer guide that includes ChatGPT-generated deals on top products, comparisons, and links to the retailer’s website for placing an order. In the future, the company intends to incorporate this feature into the Instant Checkout experience, allowing you to make purchases directly on the site.  

OpenAI has stated that user conversations are never disclosed to retailers and that the outcomes are produced organically, drawing from publicly accessible websites. Websites that wish to be included in the results must permit OpenAI’s crawlers to access their content, which can be done by adhering to the allowlisting guidelines.  

When I inquired with OpenAI about how users can be assured that the sites and products being suggested are from credible sources, I was informed that the model has been trained to avoid low-quality and spam websites, thereby helping ensure users receive the optimal user experience. During the same event, I had the chance to personally test the future, and I found it to be significantly more enjoyable than I had expected.  

My experience: 

I received an exclusive preview of the new feature, and the most notable aspect for me was its interactivity.  

For instance, during one demonstration, a user uploaded an image of Jeremy Strong from Succession and inquired about whether to find a similar turtleneck at a lower price. Although one could have performed a reverse image search on Google, this time the user was able to input a quick prompt, navigate through the refining process of “not interested,” and receive a selection of options.  

In another case, I aimed to explore suitable products for my dog. I entered the prompt, “I want to buy dog treats for a 15-pound Yorkie.” Subsequently, it posed follow-up questions, such as “Preferred treat types,” accompanied by multiple-choice answers that included:  

• Crunchy biscuits  
• Jerky bites  
• Soft chews  
• Freeze-dried  

Subsequently, it provided me with rapid feedback, allowing me to express my level of interest in the products. I was then able to select “not interested” and provide justifications, such as “too expensive” or “my dog did not favor this brand,” which helped refine the choices. Although the shopping guide was not particularly innovative, I found it beneficial and would have felt at ease purchasing one of its suggested products or, at the very least, using it as a primary reference.  

Afterward, I informed ChatGPT that I was seeking a tailored, vibrant-colored dress for a particular semi-formal event priced around $100. Among the selections it provided, some were so impressive that I captured a photo on my phone for future reference.  

However, my preferred use was to discover gifts for others. I take great pride in and enjoy gift-giving, so I was curious to see whether ChatGPT is good at finding deals that surpass my efforts. I crafted a detailed prompt. I am searching for a gift for my friend who enjoys reading and has a 12-pound dog. Yorkie has recently relocated to a new apartment, wishes to travel more in the coming year, and is fascinated by Mickey Mouse. She is 32 years old and resides in NYC.  

It required my preferred gift themes from which I chose all reading and books, dog accessories, home decor, and travel essentials. Subsequently, it presented several entertaining options, naturally due to the numerous characteristics I provided. Some product recommendations were not quite on target, particularly those that were entirely Mickey-themed. Nevertheless, some suggestions were quite considerate, with my favorite being a gift card to The Strand, a bookstore I genuinely adore in NYC, and perhaps I could not have discovered it had I not lived there.  

The Conclusion 

In the end, do I believe it will entirely replace my shopping habits? No, this is primarily because I cherish the process of finding the ideal products online, whether for someone else or for myself.  

Undoubtedly, ChatGPT compares product prices and even offers many suggestions. It could be a useful time-saving resource, particularly if I am aiming to stick to a specific budget.