Getting Started with AI Copilot

What Is Copilot

Copilot is a conversational AI assistant that understands your entire ecommerce catalog, search data, and performance metrics. You ask questions in natural language, and Copilot queries your data, synthesizes insights, and proposes actions.

Key Capabilities

• Catalog queries: "Show me products in Electronics with score below 50" or "Which products have no images?"
• Search data analysis: "What keywords drive the most impressions?" or "Which products have declining CTR?"
• Cross-reference data: "Show me products with high GMC impressions but low Search Console CTR" (merging data sources)
• Propose actions: "Queue rewrites for the 20 lowest-scoring products" or "What if we targeted 'waterproof' in our Electronics category?"
• Trigger actions: "Rewrite these 5 products" or "Start the Content Optimizer agent for products under 60"

Accessing Copilot

Location

Copilot is a chat interface in the EcomIQX dashboard. Look for the chat bubble (usually bottom-right) or a "Copilot" sidebar option. Click to open the chat window.

Tier Requirements

Copilot is Growth tier only ($249/month minimum). Starter and Free tiers do not have access.

Example Conversations

Example 1: Find High-Impact Rewrites

You: "Show me the 10 products with highest organic traffic that score below 60."

Copilot: "I found 12 products matching your criteria. Here are the top 10 by monthly organic traffic: [list]. Total potential traffic impact: 4,200 monthly sessions. Would you like me to queue rewrites for these products?"

You: "Yes, rewrite them all."

Copilot: "Queuing 10 rewrites. You will receive drafts for review in 5 minutes. Estimated score improvement: 18 points average."

Example 2: Analyze a Keyword Gap

You: "How many monthly searches are we missing in the 'wireless earbuds' category?"

Copilot: "Your wireless earbuds category targets 15 keywords with 42,000 combined monthly searches. There are 38 related keywords with 28,000 additional searches that you do not rank for. Top 5 gaps: [list with search volumes]. Would you like to see which products should target these gaps?"

You: "Show me."

Copilot: "Your 'ProAudio X3' product should target 'noise-cancelling earbuds' (3,200 searches). Your 'SportBuds Pro' should target 'waterproof wireless earbuds' (1,800 searches). Should I queue rewrites for these?"

You: "Yes, for both."

Example 3: Merge Data Sources

You: "Which products have high impressions in Google Merchant Center but declining CTR in Search Console?"

Copilot: "I found 8 products with rising GMC impressions but falling Search Console CTR. This usually means the product is being shown more in Google Shopping but the search snippet is not competitive. Products: [list]. Should I analyze their search snippets and propose title/description rewrites to improve CTR?"

You: "Yes, show me the proposals."

Example 4: Category-Level Strategy

You: "What is our keyword coverage in the Apparel category?"

Copilot: "Apparel category: 340 products, 68% have keyword coverage score above 70. 32% (109 products) score below 70. Top missing keywords: [list]. If we improve the lowest-scoring 50 products, you could capture an estimated 8,300 additional monthly searches. Want to start with those 50?"

Example 5: Launch an Agent Programmatically

You: "Start the Content Optimizer agent. Target products with score below 65 in Electronics and prioritize those with over 100 monthly organic searches."

Copilot: "Setting up Content Optimizer agent with those parameters. Found 67 eligible products. Agent will start rewrites tomorrow at 2 AM and process ~10 products per day. Estimated completion: 7 days. You will get daily status updates. Confirm?"

You: "Confirm."

How Copilot Understands Your Data

Data Access

Copilot has real-time access to:

• All product data (title, description, price, score, images, category)
• Google Search Console data (queries, impressions, clicks, position)
• Google Merchant Center data (impressions, clicks, conversion rate)
• All rewrites and their impact
• Agent activity and performance
• All scores and scoring dimensions

Query Parsing

Copilot uses natural language understanding to parse your requests. It can handle:

• Filters ("products scoring below 50", "in Electronics", "with low CTR")
• Aggregations ("total monthly searches", "average score", "count of missing images")
• Comparisons ("higher than X", "declining", "missing")
• Cross-references ("high impressions but low CTR", "products with no GSC data")

Context Awareness

Copilot understands context across conversations. If you ask "Now show me the ones in the Shoes category", it remembers you were talking about low-score products and filters accordingly.

Types of Actions Copilot Can Trigger

Rewrite Queue

"Rewrite the top 20 products by traffic that score below 60."

Copilot queues rewrites for review. You approve before they go live.

Agent Launch

"Start the Language Expander agent for products in Spanish-speaking regions."

Copilot configures and launches agents with your specified parameters.

Keyword Optimization

"Retarget these 10 products for 'wireless earbuds' instead of 'earbuds'."

Copilot queues rewrites informed by the new target keyword.

Bulk Updates

"Add 'Limited Stock' urgency copy to products in the Electronics category that have under 5 in inventory."

Copilot generates bulk rewrites with the new directive.

Limitations and When Copilot Cannot Help

Cannot Access External Data

Copilot cannot query competitor data, industry benchmarks, or external databases. It only sees your data.

Cannot Modify Feed Data Directly

Copilot can propose rewrites but cannot change prices, SKUs, availability, or other non-content feed fields. Use your product feed directly for those changes.

Cannot Browse the Web

Copilot cannot look up current pricing, stock levels, or competitor information in real-time.

Cannot Make Decisions Without Confirmation

Copilot proposes actions but waits for your confirmation before triggering them. It does not auto-execute unless you explicitly approve.

Tips for Getting the Best Results

Be Specific

Good: "Show me products in the Shoes category with score below 50 and over 1,000 monthly organic impressions."

Vague: "Show me bad products."

Use Filters Consistently

Learn the common filters (category, score range, traffic threshold, dimension) and use them consistently so Copilot learns your priorities.

Ask for Proposals Before Actions

Instead of "Rewrite these 50 products", ask "Show me rewrites for the top 5 and I will decide if I like the approach". This lets you validate before bulk actions.

Cross-Reference Data

Combine data sources: "Show me products with high GMC impressions but low Search Console CTR" is more insightful than looking at each in isolation.

Use Copilot for Discovery

Ask exploratory questions: "Which categories have the biggest keyword gaps?" Use the insights to guide your strategy.

Real-World Impact

Teams using Copilot report:

• 80% faster insight generation: What takes 2 hours to analyze in a spreadsheet takes 2 minutes in Copilot.
• Better prioritization: Copilot identifies the highest-impact opportunities, not just random low-scoring products.
• Cross-data discovery: Merging GSC + GMC + catalog data reveals insights you would miss in isolation.
• Faster iteration: You can test a hypothesis ("rewrite high-traffic, low-CTR products"), see results, and adjust in days instead of weeks.

Copilot is Growth tier and above only. Upgrade to Growth to access conversational catalog intelligence.