{
  "id": "H10",
  "url": "https://cartforagents.com/guides/ai-visibility-score",
  "title": "AI visibility score: check the denominator before you trust the number",
  "description": "Ask how a score samples products, contexts and time, handles missing data and weights components using an independent score-review checklist.",
  "answer": "Before you trust an AI visibility score, ask five things: what was sampled (products, questions, assistants and dates), what the denominator is, how missing data is handled, how the parts are weighted and what they actually contribute, and whether the score was ever checked against an outcome such as clicks or sales. A score that can't answer those describes its own method, not your visibility. We ran these checks on our own first score, and retired it.",
  "take": "We think single-number AI visibility scores do more harm than good for product brands. Ours turned out to be mostly a count of reviews, so we replaced it with plain counts of what came back and what was missing.",
  "summary": "Every AI visibility score is a fraction. Before you trust the number, find out what's on the bottom of it: which products, which questions, which assistants, which days, and what happened to everything the tool couldn't see. Here are five cases worked from public score methods, and the checklist we use.",
  "claims": [
    {
      "text": "AgentReady publishes its AI-readiness method: a weighted average of 23 always-on checks totalling 31 base weight, with full credit for a pass, half for a warn and none for a fail, an absent signal counting as a fail, and a cap at 49 when robots.txt blocks an AI search crawler. Its worked example earns 23.5 of 31, shown as 76 (grade B).",
      "source_title": "AgentReady: How the AI-readiness score is calculated",
      "source_url": "https://joinagentready.com/tools/shopify-ai-readiness-checker/how-it-works",
      "source_date": "2026-10-27",
      "method": "vendor-claim",
      "limitation": "The vendor's published method and a stand-in store. We did not run the checker."
    },
    {
      "text": "In that worked example, the llms.txt and agents.md checks each fail with a weight of 1. Excluding those two checks instead of failing them gives 23.5 of 29, or 81; scoring them as warns gives 24.5 of 31, or 79.",
      "source_title": "Cart for Agents: recomputation of a published worked example (How we checked, this page)",
      "source_url": "https://cartforagents.com/guides/ai-visibility-score#how-we-checked",
      "source_date": "2026-10-27",
      "method": "inferred",
      "limitation": "Our arithmetic on the vendor's published stand-in example, not a real store, and not a claim that the vendor's choice is wrong."
    },
    {
      "text": "AgentReady's proof page says independent studies have found no measurable ranking or citation lift from publishing an llms.txt file, and that it treats the file as machine-readable hygiene rather than a ranking lever.",
      "source_title": "AgentReady: How we practice what we check",
      "source_url": "https://joinagentready.com/proof",
      "source_date": "2026-10-27",
      "method": "vendor-claim",
      "limitation": "The vendor summarizes other testing without naming the studies on that page."
    },
    {
      "text": "usegeo says its score counts how often a brand is named and where: full points in an answer's top three, two-thirds lower down, and one-third when an answer only links to the website, averaged across the assistants. It asks three questions, each twice on each assistant.",
      "source_title": "usegeo: AI Visibility Checker",
      "source_url": "https://usegeo.app/tools/ai-visibility-checker",
      "source_date": "2026-10-27",
      "method": "vendor-claim",
      "limitation": "The vendor's description. Our worked example applies a rule of that shape to hypothetical answers; it is not usegeo's computation."
    },
    {
      "text": "FoundRate's scan page says that in its 82-store study, 87% of stores never came up once.",
      "source_title": "FoundRate: Free AI Visibility Checker for Shopify Stores",
      "source_url": "https://foundrate.co/scan",
      "source_date": "2026-10-27",
      "method": "vendor-claim",
      "limitation": "A vendor statistic. The page we read does not give the number of questions per store, assistants or dates."
    },
    {
      "text": "If a product had a fixed 10% chance of appearing in each answer, the chance of it never appearing would be 72.9% across 3 answers, 53.1% across 6, 34.9% across 10 and 12.2% across 20.",
      "source_title": "Cart for Agents: sampling arithmetic (How we checked, this page)",
      "source_url": "https://cartforagents.com/guides/ai-visibility-score#how-we-checked",
      "source_date": "2026-10-27",
      "method": "inferred",
      "limitation": "Illustrative arithmetic with an invented, constant rate and independent answers. Real answers are neither."
    },
    {
      "text": "ARC's product-data page labels counts as measured, inferred or not measured, keeps inferred scores out of its rates, dates its snapshot (12 October 2026), and marks one table as live saved checks rather than a frozen snapshot.",
      "source_title": "ARC Report: What product data do AI agents need?",
      "source_url": "https://www.arcreport.ai/answers/what-product-data-do-ai-agents-need",
      "source_date": "2026-10-27",
      "method": "vendor-claim",
      "limitation": "Vendor page. We did not verify its dataset."
    },
    {
      "text": "Recomputing our retired v0 gap score over 162 searches, review depth produced 64.7% of all points against a design weight of 25 out of 100, and reading absent ratings as zero supplied 591.4 of the review points.",
      "source_title": "Cart for Agents: AI visibility scores, audit weights and retire an invalid metric",
      "source_url": "https://cartforagents.com/research/ai-visibility-score-methodology",
      "source_date": "2026-10-27",
      "method": "observed",
      "limitation": "An audit of our own retired metric on one unpersonalized Shopify Catalog snapshot. It does not describe any vendor's score."
    },
    {
      "text": "In our 162-search snapshot, the 1,620 results are occurrences, not distinct products: product rating was present on 1,180 and absent on 440, with no explicit zeros.",
      "source_title": "Cart for Agents: Receipts cited by our guides (Shopify agent-catalog snapshot, Oct 26, 2026)",
      "source_url": "https://cartforagents.com/research/agent-shopping-research#receipts-cited-by-our-guides",
      "source_date": "2026-10-26",
      "method": "observed",
      "limitation": "One unpersonalized snapshot. Absent means unknown."
    }
  ],
  "sources": [
    {
      "title": "AgentReady: How the AI-readiness score is calculated",
      "url": "https://joinagentready.com/tools/shopify-ai-readiness-checker/how-it-works",
      "accessed": "2026-10-27"
    },
    {
      "title": "AgentReady: How we practice what we check",
      "url": "https://joinagentready.com/proof",
      "accessed": "2026-10-27"
    },
    {
      "title": "usegeo: AI Visibility Checker",
      "url": "https://usegeo.app/tools/ai-visibility-checker",
      "accessed": "2026-10-27"
    },
    {
      "title": "FoundRate: Free AI Visibility Checker for Shopify Stores",
      "url": "https://foundrate.co/scan",
      "accessed": "2026-10-27"
    },
    {
      "title": "ARC Report: What product data do AI agents need?",
      "url": "https://www.arcreport.ai/answers/what-product-data-do-ai-agents-need",
      "accessed": "2026-10-27"
    },
    {
      "title": "Cart for Agents: AI visibility scores, audit weights and retire an invalid metric",
      "url": "https://cartforagents.com/research/ai-visibility-score-methodology",
      "accessed": "2026-10-27"
    },
    {
      "title": "Cart for Agents: Receipts cited by our guides",
      "url": "https://cartforagents.com/research/agent-shopping-research#receipts-cited-by-our-guides",
      "accessed": "2026-10-27"
    }
  ],
  "limitations": [
    "Checked on 2026-10-27. Vendors change their methods; re-read a tool's method page before applying these cases to it.",
    "We read public method pages and did not run any tool. Recalculations use vendors' published stand-in examples or hypothetical answers, not real stores.",
    "The checklist is our proposal. It is not an industry standard, and passing it would not make a score predict picks, traffic or sales.",
    "Vendor figures are quoted as vendor claims, not independent measurements.",
    "Nothing here measures what any assistant tells shoppers."
  ],
  "related": [
    "https://cartforagents.com/guides/ai-product-visibility",
    "https://cartforagents.com/research/ai-visibility-score-methodology",
    "https://cartforagents.com/guides/ai-visibility-checker-accuracy",
    "https://cartforagents.com/research/missing-product-review-data",
    "https://cartforagents.com/guides/ai-product-visibility-dashboard"
  ],
  "published": "2026-10-27",
  "updated": "2026-10-27",
  "checked": "2026-10-27",
  "kind": "guide"
}