Learning Timeline
Key Insights
Interactive Refinement (The 'Tinder' Method)
The Shopping Research feature uses a 'Tinder-like' feedback loop. You must explicitly mark items as 'Not interested' or 'More like this' during the initial search phase. This real-time feedback significantly improves the relevance of the final comparison table compared to a static prompt.
Privacy & Data Source Integrity
Unlike standard browsing, your specific chat data is not shared with retailers. Additionally, the agent is fine-tuned to ignore low-quality or spammy SEO sites, strictly sourcing data from reputable, publicly available retail sites.
Plan Availability
This feature is available to all users, including those on the Free plan, with high usage limits active through the holiday season.
Prompts
Product Research Criteria
Target:
ChatGPT
whites minimalist headphones noise cancelling under $50
Step by Step
Using the Shopping Research Agent
- Log in to the ChatGPT interface.
- Locate the 'Shopping Assistant' popup overlay or navigate to the main chat input bar.
- Click the 'Try it' button on the popup (if available) or type a shopping-intent prompt directly.
- Enter specific product criteria into the chat input (e.g., color, style, feature set, price cap).
- Click the 'Get started' button to activate the autonomous research agent.
- Respond to the agent's multiple-choice clarifying questions by clicking the most relevant option tiles.
- Refine the preliminary results using the interactive rating buttons: click 'Not interested' to remove items or 'More like this' to steer the algorithm.
- Scroll down to the 'Best Overall' section to review the primary recommendation, key features, and trade-offs.
- Examine the generated 'Comparison Table' for a side-by-side view of top alternatives, best use cases, and pros/cons.
- Click the citation links within the report to verify details on external retailer websites.
