Deep Research
Let AI research and understand your products
Product information often lives across manufacturer sites, supplier pages, and documentation. Deep research lets AI search the web for information about your products when generating content, bringing in data that might not exist in your catalog yet.
Enable Use deep research on any generatable attribute and AI will search for relevant information online while generating content. Instead of working only from what's in your catalog, AI can find additional details from across the web.
Best for multi-brand retailers
Deep research works best for multi-brand retailers selling products that already have an online presence. If you only sell your own unique products, this feature may not be as relevant.
Why use deep research
Instead of manually looking up product information across manufacturer websites, supplier documentation, and technical spec sheets, let AI handle the research for you. It pulls in details like materials, certifications, measurements, and features from online sources and brings them directly into your content.
This is especially useful when your catalog has incomplete data. AI can find and fill in missing technical specifications, care instructions, or product features that are documented somewhere online but haven't made it into your system yet. The research happens automatically during generation, saving you from copying information between websites and your catalog.
How it works
When you enable deep research on an attribute, AI searches the web for relevant information when generating that attribute. It looks for product details, specifications, and other data from public websites, then uses what it finds alongside your existing catalog data to create accurate content.
The research happens automatically as part of generation. AI determines what to search for based on your instructions and the product being generated, finds relevant sources, and incorporates the information into the generated content.
Research quality depends on product visibility
Well-known products are easier to research. Nike's latest running shoes have extensive online documentation and reviews. A generic t-shirt from a local boutique might not have any web presence, making research less effective. Deep research works best with branded products or items that have an established online presence.
Enabling deep research
Deep research works with any generatable attribute. Open the attribute settings and enable Use deep research.
Once enabled, AI will search for information every time you generate that attribute in a proposal. You don't need to do anything else - it happens automatically as part of your normal generation workflow.
Deep research and proposals
Deep research works when you generate attributes through proposals. The AI searches for relevant information based on your instructions and the product context.
Viewing sources in proposals
When reviewing a proposal, you can see which sources AI used during research. This helps you understand where the information came from and verify accuracy.
Combining with instructions
Get the most from deep research by telling AI exactly what to look for. Be specific about the information you need:
Technical specifications
Find the technical specifications including materials, dimensions, and care instructionsSustainability information
Research sustainability certifications, eco-friendly materials, and environmental impactSupplier details
Look up supplier information and manufacturing details from official sourcesProduct features
Find key features and unique selling points from manufacturer documentationCare and usage
Research proper care instructions and usage guidelines for this product typeThe more specific your instruction, the better AI can focus its research on what matters for your attribute.
Common use cases
| Use Case | When to Use It |
|---|---|
| Products with limited catalog data | You have basic product information but need more details. Enable deep research to pull in additional specifications and features from manufacturer websites. |
| Technical specifications | Generate detailed specification attributes by letting AI research technical details, measurements, certifications, and materials from official sources. |
| Supplier and manufacturer information | Pull in supplier details, country of origin, and manufacturing information when it's available on supplier websites or documentation. |
| Industry-standard descriptions | Research how similar products are described online to create descriptions that match industry language and customer expectations. |
| Care and compliance information | Find care instructions, safety certifications, and compliance details from manufacturer documentation and official sources. |
Best practices
Think of deep research as a helpful bonus rather than a guarantee. Branded products and items from major manufacturers will have more information available online, while generic or unique products might not turn up much. Write your instructions so they work either way - if AI finds relevant information, great, but if not, you should still get useful content.
When writing instructions, be specific about what you want AI to research. "Find the material composition" gives AI a clear target, while "Research this product" is too vague. Better yet, write instructions with fallbacks built in:
If available online, include the manufacturer's recommended care instructions. Otherwise, provide standard care guidance for this material type.This approach ensures you get good results whether research succeeds or not.
Deep research works alongside everything else in your catalog. AI combines what it finds online with your existing product data and image analysis to create the best possible content. When you review proposals, check which sources AI used - this helps you verify accuracy and understand where the information came from.