Video tutorial
Getting Started
Navigate to the AgentIQ Product Feed application within your Shopify admin. You will land on the Product Analysis Dashboard.
Running a Product Analysis
The dashboard allows you to simulate customer searches to see how your catalog responds.
Enter Search Query: In the search bar, type a natural language query that a customer might use (e.g., "family tent with rain protection").
Configuration:
Use custom query: Ensure this box is checked to use your specific text.
Apply AI filters (Recommended): Keep this checked. The AI will automatically analyze your search intent and apply relevant filters (e.g., "Camping," "In Stock") to improve results.
Execute: Click the Run Analysis button.
Reviewing Search Results
The system will return products from your catalog that match the query.
Result Cards: Each product displays a Rank (order of appearance) and a Score (relevance rating, e.g., 0.60).
Applied Filters: The left sidebar shows which filters the AI automatically selected based on your query (e.g., Usage Type: Camping).
Note: The AI only works with native Shopify filters. Third-party or external filters are not supported.
Deep Dive: The AI Audit
To understand why a product ranked the way it did (or why it received a low score), use the AI Audit feature.
Launch Audit: Click the See AI Audit button on a specific product card.
Processing: The system will analyze product data against the search intent, extracting attributes and evaluating relevance.
Understanding Audit Metrics
The audit window presents a breakdown of the match quality:
Search Query Attributes: Break down of your query into key components (e.g., Product Type: Tent, Purpose: Family, Feature: Rain Protection).
Scores:
Overall Score: A cumulative rating of the product's match (out of 100).
Confidence: The model's certainty in its selection.
Intent Alignment: How well the product fulfills the specific user request.
Detailed Attribute Analysis & Optimization
The audit provides a section-by-section breakdown of your product data with actionable advice.
Tags: Displays which tags were successfully matched (Green) and which were expected but missing (Red).
Per-Attribute Evaluation:
Score: A rating (e.g., 1/5) for specific attributes like "Purpose: Family" or "Feature: Rain protection."
Supporting Evidence: Highlights what data did work (e.g., finding the word "tent").
Improvement Recommendations: Specific actions to take, such as:
Update product type (e.g., change "Shelter" to "Tent").
Add specific tags (e.g., "family-friendly," "waterproof").
Metadata: analyzing backend data like Vendor, Category, and Metafields.
Text Analysis: Highlights relevant phrases found in your title and description (e.g., "4-season winter tent").
Tip: If the description implies a feature (e.g., "warm and protected") but doesn't explicitly state it (e.g., "rain protection"), the AI may score it lower.
Image Analysis: The AI analyzes product images for visual context.
Recommendation: Use images that clearly depict the product's use case (e.g., a tent in a rainy environment) to improve scoring.
Optimizing Your Listings
Use the audit findings to update your product listings directly in Shopify:
Review the "Improvement Recommendations" in the audit.
Add missing tags identified by the AI.
Refine descriptions to explicitly include missing keywords (e.g., "family," "rain").
Update visuals if the AI failed to recognize the product context from images alone.