AI Readiness Score for Ecommerce

An AI readiness score is a benchmark that captures how well a retailer is set up to be discovered and recommended by AI shopping agents like ChatGPT, Google AI Mode, Perplexity, and Amazon Rufus. It typically runs on a 0-100 scale, broken down into dimensions that each measure a different part of the discovery-to-recommendation pipeline.

The score exists because search-engine optimization and AI-agent optimization are measurably different disciplines. A site that ranks well in Google's classic search results can be nearly invisible to an AI agent if its product data is thin, its feed is incomplete, or its site blocks LLM crawlers. An AI readiness score surfaces that gap before it costs you traffic.

The April 2026 Aido Lighthouse "AI Commerce Readiness Index" benchmarked 345+ retailers and found the average readiness score was 48.1 out of 100. Thirty percent scored zero on transactability. Only 2% could support end-to-end autonomous AI transactions. Those numbers tell the industry-level story: most retailers are under-indexed for the channel that is growing 393% year over year in traffic (Adobe Analytics, April 2026).

A useful AI readiness score is diagnostic, not just a single number. The most actionable scores break the measurement into distinct dimensions:

Product Mapping. Does the site expose clear product entities that AI agents can identify and track? Are product pages canonicalized, with stable URLs, unique IDs, and clean inventory signals? This dimension penalizes sites where the same SKU appears under multiple URLs or where product identity drifts between the feed and the site.

Structured Attributes. How many machine-readable attributes does each product expose? The practical benchmark is 25-40 attributes per SKU - material, dimensions, fit, compatibility, certifications, care instructions. Most retailers sit at 5-8. This dimension drives the biggest readiness swings because AI agents match queries to attributes before they fall back to keyword matching.

Attribute Context. Are the attributes described in ways a language model can reason over? "Large" is a fit attribute but "runs large - size down for a slim fit" is a reasoning signal. Listings with rich attribute context surface for nuanced queries that sparse listings miss entirely.

Product Context. Does the product page, feed, and surrounding content tell an AI agent why someone would want this product, for what use case, in what situations? This covers descriptions, use-case tags, compatibility notes, and the narrative content around the raw specs.

Scores in each dimension combine into the headline number. A site scoring 82 overall might have Structured Attributes at 95 but Product Context at 60, which is a completely different remediation plan from a site scoring 82 with the inverse mix.

The economic case for an AI readiness score is tied to where ecommerce traffic is moving. In Q1 2026, AI-referred traffic to US retailers grew 393% year-over-year (Adobe). In March 2026, AI-referred shoppers converted 42% better than typical human traffic. ChatGPT alone now drives roughly 20% of Walmart's referral traffic and 15% of Target's.

Retailers whose AI readiness is high are pulling disproportionate share of that traffic. Retailers whose readiness is low are watching it flow to competitors. The readiness score is the leading indicator: a site that improves from 40 to 80 sees measurable traffic lifts from ChatGPT, Google AI Mode, and Perplexity within weeks.

The score is also a useful internal alignment tool. SEO teams, PIM teams, and ecommerce teams often work on adjacent parts of the readiness problem without coordination. A single readiness score gives leadership a shared measurement to align those teams around. Anecdotally, the most common reaction when a retailer sees their first readiness score is "why is this not on our scorecard already?".

The practical path to a higher AI readiness score follows a predictable sequence:

1. Audit attribute depth on your top 20% of SKUs by revenue. Count the structured attributes each has. Anything under 15 is a gap. Anything under 8 is effectively invisible. This single audit surfaces most of the readiness problem for most retailers.

2. Enrich the sparse SKUs. Add missing attributes: material, dimensions, use-case tags, compatibility, certifications, care instructions. The target is 25-40 machine-readable fields per SKU.

3. Add attribute context. Rewrite sparse descriptions to include reasoning cues: "runs large", "works with X but not Y", "best for indoor use in dry climates". Language models reason over narrative context, not just keyword matches.

4. Fix feed coverage. Ensure your catalog is syndicating to Google Merchant Center (for Google AI Mode and Gemini via UCP) AND to ChatGPT via ACP. Running one but not both cuts your surface coverage in half.

5. Unblock AI crawlers. Check your robots.txt allows GPTBot, Google-Extended, PerplexityBot, ClaudeBot, and Anthropic-AI. Add an llms.txt file describing your catalog structure and canonical endpoints.

6. Measure weekly. A readiness score is a leading indicator only if you watch it. Run the same queries across ChatGPT, Google AI Mode, and Perplexity every week. See which products surface as product cards, which as text mentions, and which disappear.

Paz.ai, an agentic commerce optimization platform, offers a free AI Readiness Report at paz.ai/ai-readiness that scores any product URL across the four dimensions above in under a minute.