Is Your Product Catalog Ready for AI Shopping Agents? A Self-Assessment
The shift to agentic commerce is not arriving with an announcement or a deadline. It is accumulating, one agent-driven product search at a time, one Gemini product recommendation, one Rufus query, one browser AI shopping session. By the time most retailers treat it as urgent, the first-mover window will have closed. This self-assessment gives you the framework to measure where your catalog actually stands today, across the six dimensions that determine agent visibility, and a prioritized action plan based on where your biggest gaps are.
How to Use This Assessment
Work through each of the six dimensions below. For each, follow the measurement instructions to produce a score for your catalog. Record your scores honestly. A clear view of where you are is more valuable than a comfortable estimate. The action plan at the end sequences the highest-impact improvements based on your score profile.
Use Your Actual Data, Not Your Best Category
Run this assessment on a representative sample of your catalog, not your hero products or your best-maintained category. Pull 100 SKUs from your top-3 revenue categories, including recent additions. The score on those SKUs reflects your actual agent visibility, not your potential. If you assess only your cleanest data, you will underestimate the gap.
The six readiness dimensions
Attribute completeness
Do the fields exist where agents need them?
Value precision
Are values specific and machine-queryable?
Schema implementation
Can browser agents and crawlers read your PDPs?
Real-time accuracy
Do price and availability stay current everywhere?
GTIN coverage
Can knowledge graphs entity-match your products confidently?
Taxonomy depth
Are products classified deeply enough to qualify for precise queries?
Dimension 1: Attribute Completeness
This is the most commercially impactful dimension. Every missing purchase-criteria attribute is a complete exclusion from agent queries that include that criterion.
How to measure
- Step 1 — For your primary category, list the 10 attributes a shopper is most likely to specify when buying in that category. For hiking jackets: waterproof, weight, packable, material, gender, fit, size range, hood type, sustainability cert, activity type.
- Step 2 — Pull 100 representative SKUs from that category.
- Step 3 — For each of the 10 attributes, count what percentage of the 100 SKUs have that attribute populated as a structured field, not mentioned in description text, but in a dedicated attribute field.
- Step 4 — Average the 10 percentages to produce your Attribute Completeness Score.
Not Ready
Attribute Completeness score below 50%.
Progressing
Attribute Completeness score between 50% and 79%.
Agent-Ready
Attribute Completeness score at or above 80%.
Dimension 2: Value Precision
Completeness without precision is not agent-readable. A weight attribute containing “very lightweight” is structurally present but functionally absent. An agent cannot execute weight < 500g against a text string.
How to measure
- Step 1 — From the same 100 SKUs, pull all populated attribute values for your top-5 numeric or categorical attributes such as weight, dimensions, waterproof rating, and similar fields.
- Step 2 — For each attribute value, classify as Precise (specific number with unit: “490g”) or Imprecise (descriptive: “lightweight,” “approximately 500g,” “very water resistant”).
- Step 3 — Calculate the percentage of populated attribute values that are Precise.
Not Ready
Value Precision score below 40%.
Progressing
Value Precision score between 40% and 74%.
Agent-Ready
Value Precision score at or above 75%.
Dimension 3: Schema.org Implementation
Schema.org Product markup is the data layer that browser agents, AI crawlers, and Google’s indexing system read independently of your product database. Without it, products on your DTC site may be invisible to agents that read PDPs rather than Shopping feeds.
How to measure
- Step 1 — Take 20 product URLs from your DTC website and run each through Google’s Rich Results Test.
- Step 2 — For each URL, check: (a) Is Product schema present? (b) Are Offer fields (price, availability) present? (c) Is aggregateRating present? (d) Are additionalProperty fields present for key attributes?
- Step 3 — Score each URL: 0 = no schema; 1 = basic schema only; 2 = Product + Offer; 3 = Product + Offer + Rating; 4 = full schema including additionalProperty. Average across 20 URLs and divide by 4 to get a percentage score.
| Score Range | Interpretation |
|---|---|
| Below 30% | Not Ready |
| 30%–69% | Progressing |
| 70% and above | Agent-Ready |
Dimension 4: Real-Time Data Accuracy
Agents making purchase decisions need to trust that your data is current. An agent that recommends a product that turns out to be out of stock, or priced differently than stated, has failed its user. Over time, platforms learn to down-weight merchants with accuracy problems.
How to measure
- Step 1 — Check your Google Merchant Center diagnostics: how many active price mismatch or availability mismatch errors do you have right now?
- Step 2 — Measure your price change propagation latency: when you change a price on your website, how long does it take to appear in your Merchant Center feed and in your schema markup? Measure this for a real price change event.
- Step 3 — Check your schema.org price and availability fields against your actual website right now for 10 random products. How many match exactly?
Not Ready
Propagation latency greater than 24 hours.
Progressing
Propagation latency between 2 and 24 hours.
Agent-Ready
Propagation latency under 2 hours.
Dimension 5: GTIN and Entity Coverage
GTINs enable entity matching, the process by which agents and knowledge graphs map your product to a known product entity. Entity-matched products have higher confidence scores in Shopping Graph queries and are eligible for cross-merchant comparison formats that non-entity-matched products cannot access.
How to measure
- Step 1 — Export your Merchant Center product feed and filter for the GTIN field. Count how many products have a valid GTIN populated versus how many have it empty or as a placeholder.
- Step 2 — If you are on Amazon, check your ASIN records for GTIN or EAN population in the relevant inventory reports.
- Step 3 — For any branded products without GTINs, verify whether the product has a GS1-registered GTIN that you are not submitting.
Not Ready
GTIN or Entity Coverage below 50%.
Progressing
GTIN or Entity Coverage between 50% and 79%.
Agent-Ready
GTIN or Entity Coverage at or above 80%.
Dimension 6: Taxonomy Depth and Accuracy
Agents use your category classification to determine which queries your product is eligible to appear for. Products classified at a parent category level miss query eligibility for subcategory-specific searches. Products classified incorrectly may appear for queries they cannot satisfy, a worse outcome than not appearing at all.
How to measure
- Step 1 — Pull your Google product_type values for your top category. Count how many are classified to 3+ levels deep versus 1–2 levels.
- Step 2 — Check 10 random products against Google’s product taxonomy tool. Are they mapped to the most specific applicable category node, or to a parent node?
- Step 3 — On Amazon, check 10 random ASINs in Seller Central. Are they mapped to the most specific browse node available for their product type?
Not Ready
Taxonomy Depth below 40%.
Progressing
Taxonomy Depth between 40% and 74%.
Agent-Ready
Taxonomy Depth at or above 75%.
Reading Your Score Profile and Prioritizing Action
Once you have scores across all six dimensions, your profile tells you where to focus. The prioritization logic:
| If Your Lowest Score Is In... | Prioritize This Action First | Why |
|---|---|---|
| Attribute Completeness | Category criteria mapping + structured attribute completion sprint across top-revenue SKUs | Every missing purchase-criteria attribute is a complete query exclusion. This has the highest direct commercial impact. |
| Value Precision | Audit and rewrite all descriptive attribute values for top-10 filterable attributes to unit-based, numeric equivalents | Imprecise values are functionally absent for agent filtering. Quick fix per attribute; high impact per fix. |
| Schema Implementation | Implement or audit Product + Offer + aggregateRating + additionalProperty schema across full DTC catalog | Without schema, your DTC products may be invisible to browser agents and AI crawlers, a growing share of discovery traffic. |
| Real-Time Accuracy | Implement Content API for real-time price and availability feed sync; audit schema generation to be dynamic | Stale data generates trust failures that compound into merchant reliability down-weighting over time. |
| GTIN Coverage | GTIN sourcing and submission sprint for all branded products without valid GTINs | GTIN gaps are one of the highest-single-impact data tasks for Google Shopping agent visibility. |
| Taxonomy Depth | Re-classify top-revenue products to most specific category nodes across Google and Amazon | Broad classification reduces query eligibility scope, easy to fix with a one-time audit pass. |
Find the lowest score
The bottleneck usually determines the next best action more clearly than the average score.
Fix what changes eligibility first
Completeness, precision, and taxonomy usually move agent visibility fastest.
Then improve trust layers
Schema, freshness, and GTIN coverage raise confidence and reach once basics are in place.
Velou’s Agentic Readiness Audit
Commerce-1 automates this entire assessment, running structured attribute completeness rates, precision scoring, schema validation, GTIN coverage analysis, and taxonomy depth checks across your full catalog in a single pass. The output is a SKU-level readiness report with commercial impact estimates for each gap: “closing the weight attribute gap in your Jackets category would make 340 SKUs eligible for an estimated 15,000 additional agent query matches per month.”
If you want to move from a self-assessment to a full catalog-scale readiness report, this is what we do.
Get a full agentic readiness audit, not just a self-assessment score
Commerce-1 runs the full assessment across every SKU in your catalog with commercial impact estimates.
Request a demo
.png)
.png)