16 signals. Five pillars. One evidence-backed readiness score.
CEI separates readiness into the structural conditions AI systems need before they can recommend products accurately, then combines them into a single 0 to 100 Commerce Eligibility Index™.
The conditions AI needs to read, compare, trust, and recommend.
Foundation
Is the catalog data clean and complete enough for AI to read?
Differentiation
Can AI tell products apart, or do they collapse into generic clusters?
Retrieval
Can AI find the right products for real shopping queries?
Integrity
Are the claims in your catalog internally consistent and verifiable?
Authority
Does the broader internet give AI reasons to trust and cite your brand?
The 16 signals, in plain terms.
Each pillar resolves into specific, evidence-backed signals. The names below are simply what each one checks.
- Attribute completeness: are the fields that matter filled in?
- Terminology consistency: is the same thing named the same way?
- Visual quality: are images clear, described, and on-point?
- Video discoverability: is there product video to draw on?
- Differentiation strength: are descriptions genuinely distinct?
- Tradeoff density: can AI see meaningful product-vs-product tradeoffs?
- Cluster separability: do products group cleanly into categories?
- Retrieval quality: does the right product surface for a real query?
- Constraint satisfaction: are price, size, and spec constraints honored?
- Claim consistency: do the product's claims agree with each other?
- Evidence coverage: are claims backed by traceable evidence?
- Owned-review sentiment: what do reviews on your own surfaces say?
- Independent-review sentiment: what do third-party review platforms say?
- Brand-mention density: how visible is the brand in independent discussion?
- AI-assistant discoverability: how often do assistants recommend you?
- Video authority: is there credible third-party video about the brand?
Five pillars. 16 signals. One 0 to 100 index. What we check is above; how we weight and score it is the part we keep proprietary.
The score is useful because the evidence is inspectable.
CEI combines signal scores into a 0 to 100 readiness score. Missing or unmeasurable signals are handled explicitly, not hidden as fake zeros.
High readiness across the measured signal set.
Usable foundation with meaningful improvement opportunities.
Catalog gaps are likely affecting AI discovery or accuracy.
Serious readiness gaps or insufficient evidence to score.
A measurement, not a model of one.
We do not estimate what AI sees. We crawl your catalog the way AI does, generate over 1,000 LLM-generated shopping queries from your own product vocabulary (constraint searches, comparisons, use-case and compatibility questions), and test the live assistants directly: how often ChatGPT, Claude, Gemini, and Perplexity actually recommend you when shoppers ask real product questions. Because AI needs every dimension to work, CEI is weighted so a strong pillar cannot paper over a weak one. The lowest pillar pulls the score.
Before any pillar can score, CEI runs the gate checks an AI agent runs first: bot accessibility (can AI crawlers reach your pages at all?), schema markup (is product data machine-readable?), content structure (is the page parseable without a browser?), FAQ presence, citation optimization, Universal Commerce Protocol (UCP) readiness (the agentic-commerce manifest AI agents look for), and Agentic Commerce Protocol (ACP) readiness. Video today is measured by presence and coverage; deeper video-content grading, such as transcripts and caption alignment, is in development.
Every score, read in context.
A single number means little without a baseline. That is why CatalogSignal runs the CEI Benchmark, a monthly longitudinal study of AI catalog readiness, so your CEI can be read against your category and tracked as AI shopping shifts.
The benchmark spans ten consumer-retail verticals, Beauty & Cosmetics, Consumer Electronics, Drugstore & Discount Retail, Fashion & Apparel, General Merchandise & Marketplaces, Home & Furniture, Home Improvement, Jewelry & Accessories, Pet Supplies, and Sporting Goods & Outdoor, with new verticals added over time. It establishes per-vertical baselines and tracks how readiness moves, so your score is a position in a market, not a number in a vacuum.
No. It starts where SEO stops.
Good SEO makes your pages findable, and a lot of that hygiene helps AI too. But ranking was never built to answer the question an assistant actually asks: can I confirm this product fits, trust its claims, and recommend it without getting it wrong? That lives in your product data, SKU by SKU, and in the accuracy of what assistants say about you. CatalogSignal measures and fixes that layer. It builds on your SEO; it does not replace it.
Get a readiness view for your own catalog.
A baseline CEI assessment turns the framework into a score, evidence, and a prioritized fix plan.
Request a CEI assessment