ChatGPT, Alışveriş Araştırma Asistanı ile Genişliyor: Ürün Keşfini Kullanıcı Tercihlerine Göre Özelleştiriyor

In Brief

OpenAI rolled out Shopping Research, a shopping assistant within ChatGPT that builds personalized buyer guides by scanning trusted retail sources and asking targeted questions about preferences.

AI research organization OpenAI announced the introduction of Shopping Research, an interactive shopping assistant integrated into ChatGPT that compiles tailored buyer guides by reviewing reputable retail sources and gathering information about user preferences.

The feature is designed to handle product research directly within ChatGPT, allowing users to describe what they need rather than checking multiple websites. Whether the request involves identifying the quietest cordless vacuum for a small apartment, comparing several bicycle models, or finding a suitable gift for a young child interested in art, Shopping Research generates a structured guide to support the decision-making process. It identifies relevant questions, conducts broad online research, assesses reliable sources, and utilizes ChatGPT’s existing understanding from prior conversations and memory to produce a personalized guide within minutes.

The rollout begins on both mobile and web for logged-in ChatGPT users across Free, Go, Plus, and Pro plans. To support holiday-season demand, usage limits will remain minimal across all plans during this period.

Large numbers of people already rely on ChatGPT to explore, evaluate, and compare products, especially when navigating many options and determining which ones meet specific needs, budgets, or preferences

Shopping Research is intended to assist with more complex decisions by turning product evaluation into an interactive dialogue. It gathers precise information, retrieves current data from credible sources, and returns refined recommendations that can be adjusted according to user feedback

The tool shows particular effectiveness in categories that involve extensive specifications, including electronics, household tools, beauty products, kitchen appliances, and outdoor equipment.

For straightforward inquiries such as checking a price or confirming the presence of a feature, standard ChatGPT responses remain sufficient. However, when more comprehensive analysis is required—such as comparing trade-offs or narrowing choices—Shopping Research provides a more detailed and thoroughly researched output.

The feature is also being added to ChatGPT Pulse, available to ChatGPT Pro users. When appropriate, Pulse may proactively surface personalized buyer guides based on earlier discussions. For instance, ongoing conversations about e-bikes could lead to a future Pulse card recommending relevant accessories.

Shopping Research Enhances Product Search With Tailored Recommendations And Transparent Citations

In order to begin using the assistant, users can ask a shopping-related question, after which ChatGPT may automatically propose using Shopping Research. The session can be started by selecting that prompt or by choosing the “shopping research” option from the (+) menu.

At the start of a session, ChatGPT presents a visual interface designed for discussing product options and providing feedback that shapes the research process. Users describe what they are trying to find and respond to follow-up questions about aspects such as budget, intended use, or preferred features. When memory is enabled, the assistant can further tailor the research; for instance, if previous conversations indicate an interest in gaming, that context may be considered when suggesting a new laptop.

During the research phase, the system scans the internet for current information, including pricing, availability, reviews, specifications, and images. It then returns product options progressively as the search develops. Users can steer the direction of the research by engaging with the results, dismissing items that are not relevant, or requesting more suggestions similar to those they prefer. This ongoing input allows the recommendations to adapt in real time.

After several minutes, the assistant delivers a tailored buyer’s guide summarizing leading product options, their distinguishing factors, relevant trade-offs, and the most up-to-date information drawn from reputable retailers. This provides a consolidated overview that would typically require extensive manual comparison. If the user chooses to purchase an item, they can follow a link to the retailer’s website. Future updates will allow direct purchases within ChatGPT for merchants participating in Instant Checkout where available.

The system is powered by a version of GPT-5 mini that has undergone reinforcement learning specifically targeted at shopping-related tasks. It has been trained to evaluate trusted sources, provide citations, and compile information from multiple references to produce comprehensive research. The experience is built to remain interactive, adjusting continuously to new constraints and user feedback to create a response that is both detailed and contextually aligned.

Shopping Research is designed with transparency and user trust in mind. Conversations are not shared with retailers, and all results rely on publicly accessible retail sites, with the assistant reading product pages directly, citing sources, and avoiding unreliable or low-quality material.

Although the model performs well in internal assessments focused on accurate citation of product details, it is not infallible. It may occasionally misstate information such as pricing or availability, and users are encouraged to verify final details on the merchant’s website.

This release represents an early step in simplifying product discovery. Over time, ChatGPT is expected to refine its ability to understand user preferences, broaden category coverage, and introduce more intuitive methods for exploring and comparing products.