Part One: My Store
My Store is the catalog health and compliance dashboard in AgentIQ. It gives merchants a real-time view of how well their product catalog is structured for discovery by AI shopping agents. Every metric, warning, and recommendation on this page traces directly to whether a listing will win or lose a recommendation when a shopper queries ChatGPT, Gemini, Perplexity, or a similar agent.
Figure 1 — My Store dashboard showing the Catalog Health Tracker and score breakdown
1. Catalog Health Tracker
The Catalog Health Tracker is the top-level summary card. It shows the total product and collection count, a full-width progress bar segmented by health tier, and the distribution of products across four tiers.
Catalog stats | 134 Products · 3 Collections (example store) |
Refresh scores | Triggers a re-evaluation of all scores against current listing data |
Export Report | Exports a full catalog health report for offline review or sharing |
Catalog Guardrails | Opens guardrail configuration settings (see Section 5) |
1.1 Health Tiers
Every product in the catalog is assigned one of four health tiers. The progress bar on the tracker card shows their relative proportions at a glance.
Tier | Example % | What it means |
Excellent | 58% | Listing is well-structured; AI agents can confidently recommend this product. |
Mediocre | 40% | Listing has gaps that reduce AI agent confidence. Action recommended. |
Poor | 1% | Listing has significant metadata deficiencies. Agent discovery is unreliable. |
Bad | 1% | Listing is critically under-optimized. Agents are unlikely to surface it. |
2. Catalog Health Breakdown
Below the tracker, three score cards decompose the overall health rating into specific signal categories.
Content Quality Score (80%) | Evaluates image count, description length, and title formatting across all products. |
Product Score (98%) | Evaluates category depth and attribute completeness. |
Tags Score (93%) | Evaluates standardized tag usage and search optimization. |
Note: A high Product Score does not offset a low Content Quality Score. AI agents evaluate all signals simultaneously — a weak description can cause a miss even when category and tags are perfect. |
Figure 2 — Catalog health breakdown scores and the expanded Important Warnings panel
3. Important Warnings Detected
The Important Warnings section is a collapsible panel that surfaces the most impactful issues across the catalog. Each warning card shows how many products are affected, why it matters for AI agent performance, and a direct action link.
4. Product Optimization Opportunities
Below the warnings, the Product Optimization Opportunities table provides a per-product view of all compliance issues. This is the main triage surface for individual product fixes.
Figure 3 — Product Optimization Opportunities table with health badges, alerts, and Fix Now actions
4.1 Table Columns
Product | Product thumbnail, name, and SKU handle. |
Health | Color-coded tier badge: Excellent (green), Mediocre (amber), Poor (red). |
Alerts | Count of active issues on this product. |
Suggested guardrails | Specific issue labels (e.g., Description Too Short, Over Tagged, Deepen Category). |
Action | "Fix Now" button — opens the Product Listing detail page for that product. |
Status | Shopify publish status: Published or Draft. |
4.2 Filtering the Table
Two dropdown filters sit above the table:
Status filter — Published, Draft, or All. Defaults to Published.
Health filter — All health, Excellent, Mediocre, Poor, or Bad. Use this to triage the worst-performing products first.
A search field at the top allows filtering by product title across the full 134-product catalog.
5. Catalog Guardrails
The Catalog Guardrails button (gear icon, top-right) opens a configuration panel for setting the thresholds and rules that determine when a product triggers a warning. Rules can be customized per merchant to reflect their specific catalog requirements and quality standards. (See Catalog guardrails article in the knowledge center)
Note: Guardrail customization is an advanced setting. Changing thresholds affects how the health scores and warnings are calculated for all products in the catalog. |
Part Two: Product Listing
The Product Listing page is the individual product optimization workspace. It opens when a merchant clicks "Fix Now" from the Product Optimization Opportunities table. Here, merchants can inspect every aspect of a single product's AI-readiness, run Prompt Analysis to simulate how an AI agent would evaluate the listing, and act on AI-generated suggestions to improve it.
Figure 4 — Product Listing page for "Oakley Flight Deck M Goggles" showing the header, performance metrics, and Product Pulse
6. Header and Performance Metrics
6.1 Performance Metrics Bar
Four key metrics are displayed in a row beneath the header. These reflect actual customer behavior data from your Shopify store for the selected date range.
PDP Sessions — number of product detail page views
Checkout Rate — share of PDP sessions that reached checkout
Conversion — share of PDP sessions that completed a purchase
Orders — total orders placed for this product
8. Organization Panel
The Organization panel occupies the left two-thirds of the Product Listing page. It is structured into five sub-sections covering all fields that AI agents evaluate when deciding whether to recommend the product.
Figure 5 — Organization panel showing vendor, product type, product options, category taxonomy, and the Prompts Analysis sidebar
8.1 Core Metadata
The top of the Organization panel lists the foundational product metadata fields:
Vendor — brand or supplier name (e.g., Oakley)
Product Type — category classification from Shopify (e.g., Goggles)
Handle — the URL slug used to identify the product (e.g., oakley-flight-deck-m-goggles)
Published — the date the product was first published on the store
8.2 Product Options
Product Options shows the variant dimensions defined for the product (e.g., Color: Black). These are pulled directly from Shopify and inform AI agent understanding of available configurations.
8.3 Category
The Category field shows the full Shopify taxonomy path assigned to the product. A deep, accurate taxonomy path is one of the strongest signals for AI agent routing.
Example path:
Sporting Goods > Outdoor Recreation > Winter Sports & Activities > Skiing & Snowboarding > Ski & Snowboard Goggles
Two tools assist with category accuracy:
Search Shopify taxonomy categories — free-text search to find and assign a more precise taxonomy node
"Suggest a better category" link — triggers an AI-generated category recommendation based on the product's current content
Note: A deeper taxonomy path increases the precision of AI agent routing. "Ski & Snowboard Goggles" is significantly more specific than "Sporting Goods" and directly improves match quality for intent-driven queries. |
8.4 Metafields
The Metafields section shows structured data fields attached to the product. The panel displays:
A suggestion / gap summary (e.g., "0 suggestions · 1 gap — Missing metadata for agentic discoverability")
Global metafields: Title Tag and Description Tag
The gap count indicates missing fields that AI agents rely on for structured attribute extraction. Filling these gaps directly improves agent confidence when evaluating the product against a query.
Figure 6 — Organization panel with full category taxonomy, metafields, and Prompts Analysis results
9. Prompts Analysis
Prompts Analysis is the right-side panel of the Product Listing page. It is the core diagnostic tool in AgentIQ: it lets merchants simulate the exact prompt an AI shopping agent would receive and see whether the current listing would be recommended in response.
9.1 How to Use Prompts Analysis
To run an analysis:
Type a natural-language shopper query in the "Add prompt..." field (e.g., "I'm looking for a snowboard goggle for foggy conditions and sun all mountain skiing")
Press the Run button (black button with play icon) to submit
Wait for the AI to analyze the listing — a progress bar shows "AI is analyzing your listing... 50%"
Review the Analysis Results panel that appears once complete
Figure 7 — Prompts Analysis panel with a prompt entered and ready to run
9.2 Analysis Results
The Analysis Results panel provides a structured breakdown of exactly why the product would or would not be returned for the entered query. It contains:
A narrative summary explaining the product's fit (or mismatch) against the query intent
Issue list (red triangle icons): specific listing weaknesses that caused the product to score poorly against the query, such as:
Title does not explicitly include high-intent query terms
Recommended metafields are unset, so the agentic system has less structured attribute data than competitor listings
Tags include irrelevant terms that may dilute relevance
Description is too generic; does not clearly tie the product to the query scenario
Media alt text is generic and misses visible product signals
Recommendations (green checkmark icons): specific actions to improve the listing for this query type
Note: Analysis Results are prompt-specific. Running different prompts on the same product will surface different issues and recommendations. Use this to stress-test your listing against the full range of queries your customers might send to an AI agent. |
9.3 Prompt Library
The icon to the right of the "+" button shows the total number of prompts saved in the Prompt Library (e.g., "69"). Clicking it opens the Prompt Library where pre-built and saved prompts can be selected and re-used across products. (check out our article that explains Prompt libraries)
9.4 Filters
The Filters section below the prompt field scopes the competitive analysis to a specific context. Active filters affect how the AI evaluates the listing relative to competitor products.
Location | Filter competitor comparison to a specific country market (e.g., United States, United Kingdom, Japan). |
Price range | Scope the competitive set to products within a Min–Max price band. |
My shop only | Toggle to limit analysis to your own catalog only (no cross-store comparison). |
In stock only | Toggle to include only in-stock products in the competitive set. |
Include second hand | Toggle to include used/refurbished listings in the competitive set. |
Category | Refine to specific product categories (e.g., Running Shoes, Trail Running, Athletic Wear). |
A "How does analysis work?" link at the bottom of the Filters panel opens an explainer on the methodology behind Prompts Analysis scoring.
10. AI Suggestions
AI Suggestions appear inline within the Organization panel whenever AgentIQ detects a specific, actionable improvement. Each suggestion card contains:
Impact label — Medium Impact or High Impact, indicating the expected improvement to agentic discoverability
Issue description — an explanation of what the current field contains and why it is under-performing
Detected attributes — a list of visual or descriptive signals already present in the product data
Competitive context — a note on how competitor listings with richer attributes are likely outperforming this one
Suggested replacement text — a pre-written field value the merchant can review
Accept / Dismiss buttons — one click to apply the suggestion directly to the product, or dismiss to skip it
Figure 9 — AI Suggestion card for image alt text with detected attributes, competitor context, and an editable suggested replacement
Example — alt text suggestion for Oakley Flight Deck M Goggles:
AI SUGGESTION • Medium Impact Suggested alt text: Oakley Flight Deck M ski and snowboard goggles with black frameless frame, mirrored blue-green Prizm lens, and black strap with white Oakley logo Why: The current alt text is too generic. A descriptive alt text gives the agentic system stronger visual signals for snowboard/ski use, frameless design, lens appearance, and brand-specific details. |
Note: Accepting a suggestion writes the change directly to the product in Shopify. Review the suggested text before accepting, especially for fields like Title Tag that affect how the product appears in search results. |
11. Quick Reference
Common tasks and where to find them:
Task | Where |
See overall catalog AI-readiness | My Store → Catalog Health Tracker |
Find worst-performing products | My Store → Product Optimization Opportunities → filter Health = Poor/Bad |
See which issue affects the most products | My Store → Important Warnings Detected (expand) |
Fix a specific product | Product Optimization Opportunities → Fix Now → Product Listing page |
Check category accuracy | Product Listing → Organization → Category section |
Fill missing metafields | Product Listing → Organization → Metafields section |
Test a product against a shopper query | Product Listing → Prompts Analysis → enter prompt → Run |
See why a product was not recommended | Product Listing → Prompts Analysis → Analysis Results (issue list) |
Apply an AI-generated field improvement | Product Listing → AI Suggestion card → Accept |
Filter Prompts Analysis by market | Product Listing → Prompts Analysis → Filters → Location |
Re-run all catalog scores | My Store → Catalog Health Tracker → Refresh scores |
