Agentic Commerce Visibility Report

Purpose

The merchant-facing summary skill. For a given Shopify store, runs 100+ relevant shopping queries against 5 AI agents in parallel, parses recommendations, and computes the store's AI Visibility Score — the percentage of queries where at least one of the store's products is mentioned, broken down per agent and per vertical.

Use this skill to answer: "If a buyer asks ChatGPT or Gemini for the best [vertical] product, how often does my store actually come up?"

It is the merchant-readable output of the [agentic-commerce-query-shopping-agents](../agentic-commerce-query-shopping-agents/SKILL.md) primitive, batched at scale.

Prerequisites

• Authenticated Shopify CLI session: shopify auth login --store
• API scopes: read_products
• API keys for at least one shopping agent (see [agentic-commerce-query-shopping-agents](../agentic-commerce-query-shopping-agents/SKILL.md))

Parameters

| Parameter | Type | Required | Default | Description | |-----------|------|----------|---------|-------------| | store | string | yes | — | Store domain (e.g., mystore.myshopify.com) | | vertical | string | yes | — | Vertical to test against. One of: apparel, beauty, home, electronics, fitness, food, pets, baby, outdoor, gifts | | querycount | integer | no | 100 | Number of shopping queries to run (more = more reliable score, longer runtime) | | agents | array | no | all-configured | Subset of agents to query | | format | string | no | human | Output format: human or json | | querysource | string | no | builtin | Where to source queries: builtin (curated per-vertical queries), or path:./queries.txt (one query per line) |

Workflow Steps

1. Determine the store's public domain: query 1 product, parse its onlineStorePreviewUrl to extract the public domain (e.g., mystore.com not mystore.myshopify.com). AI agents recommend by public URL.

1. Sample queries:

• If query_source: builtin, load the curated query bank for the vertical (e.g., 200 apparel queries spanning intents: best-under-price, gift-for-persona, sustainable, compare). Sample query_count queries.
• If query_source: path:./queries.txt, read line by line.

1. For each query, invoke [agentic-commerce-query-shopping-agents](../agentic-commerce-query-shopping-agents/SKILL.md) to capture cross-agent recommendations. Run in batches of 10 to manage rate limits.

1. For each captured response, check if the store domain appears in any of the recommendations: parse merchant_domain from the mention list. Track per-agent, per-query.

1. Compute aggregate metrics:

• Per-agent visibility: % of queries where store appears at least once
• Per-agent first-mention rank: median position when store appears
• Cross-agent visibility: % of queries where store appears in at least 2 of 5 agents (the "consensus" metric)
• Vertical-relative score: visibility vs the median of all stores in this vertical (requires Benchmark data)

1. Output report:

``` AI Visibility Report — mystore.com (apparel) ============================================ Queries tested: 100 × 5 agents = 500 capture events

Per-agent visibility:

• ChatGPT: 34% (median rank when seen: 4.2)
• Gemini: 28% (median rank: 5.1)
• Claude: 19% (median rank: 6.0)
• Mistral: 8% (median rank: 7.5)
• DeepSeek: 12% (median rank: 6.8)

Cross-agent consensus (≥2 agents): 22% Vertical median (all apparel): 28% Your position: below median (-6 pts)

Top 3 queries where you appear: ... Top 3 queries where you do NOT appear (you SHOULD): ... ```

Safety

> Read-only. Sends data to external AI APIs. > Costs: $5-15 per full run (100 queries × 5 agents at ~$0.003-0.01/agent/query). > Run weekly or monthly, not daily. The metric does not move that fast.

Companion skills

After this report, the merchant usually wants to:

1. Run [agentic-commerce-catalog-audit](../../catalog-audit/agentic-commerce-catalog-audit/SKILL.md) to find the lowest-scoring products that are likely the cause of low visibility
2. Fix the top issues via [agentic-commerce-generate-ai-descriptions](../../catalog-fix/agentic-commerce-generate-ai-descriptions/SKILL.md) + [agentic-commerce-generate-alt-text](../../catalog-fix/agentic-commerce-generate-alt-text/SKILL.md) + [agentic-commerce-inject-jsonld](../../catalog-fix/agentic-commerce-inject-jsonld/SKILL.md)
3. Re-run this report 4 weeks later to measure visibility uplift

Reference

The query bank per vertical is published at agentic-commerce-benchmarks (queries/.txt). The benchmarks dataset (cross-store visibility distribution per vertical) is what enables the "vertical median" comparison.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: commerce-agentic
• Source: commerce-agentic/agentic-commerce-skills
• License: MIT

Install and usage instructions live in the source repository linked above.