Agentic commerce statistics: measured, dated, cited

Measured numbers (6 November 2026)

Each figure links to its own anchor, so you can cite one number. All of them are counted from the 10,750 store rows in snapshot 2026-11-06 of ARC's open dataset.

1. 99.5%

   Stores with a completed check

   10,698 of 10,750 stores listed in the snapshot. Opted-out stores are left out. A check that did not finish is not a result and stays out of every rate below. Checks ran from 2026-10-08 to 2026-11-06 (UTC).

   Snapshot 2026-11-06, dated 6 November 2026. #stores-checked

2. 67.5%

   Returned an agent checkout link

   7,220 of 10,698 completed store checks (turned-away checks count as not reached, as on /data and the homepage). ARC's automated HTTP check asked the store's agent catalog for a product and got a checkout link back. A link is not an order: ARC stops before payment.

   Snapshot 2026-11-06, dated 6 November 2026. #checkout-link

3. 71.4%

   Agent catalog returned the product ARC searched for

   7,635 of 10,698 completed store checks. The first step before a checkout link.

   Snapshot 2026-11-06, dated 6 November 2026. #found-products

4. 4.5%

   Turned ARC's automated check away

   477 of 10,698 completed store checks. A response to ARC's checker, not a finding against the store.

   Snapshot 2026-11-06, dated 6 November 2026. #turned-away

5. 1.8%

   Block AI crawlers in robots.txt

   188 of 10,281 completed checks with a readable robots.txt. An unreadable robots file is not a block and is left out of the denominator.

   Snapshot 2026-11-06, dated 6 November 2026. #robots-ai-crawlers

6. 1.1%

   Block AI answer agents in robots.txt

   116 of 10,281 completed checks with a readable robots.txt. Answer agents are the user-triggered fetchers that assistants use to read a page for a shopper.

   Snapshot 2026-11-06, dated 6 November 2026. #robots-answer-agents

7. 91.9%

   Publish a UCP file at /.well-known/ucp

   6,740 of 7,333 completed checks where the file check got an answer. HTTP 200 JSON with a version and services or capabilities. A file check that was not recorded or was turned away is left out.

   Snapshot 2026-11-06, dated 6 November 2026. #ucp-file

8. 76.6%

   UCP profile valid against the published schema

   8,193 of 10,698 completed store checks. Valid means the profile parses and matches the schema. It is not a promise of a checkout link.

   Snapshot 2026-11-06, dated 6 November 2026. #ucp-valid

9. 88.1%

   UCP-valid stores that also returned a checkout link

   7,220 of 8,193 stores with a valid UCP profile. Declared, valid and returning a checkout link are different counts.

   Snapshot 2026-11-06, dated 6 November 2026. #ucp-valid-checkout

10. 75.9%

    Completed checks on Shopify

    8,115 of 10,698 completed store checks. Platform detection is from public signals. ARC's directory leans to Shopify, so quote platform rates separately.

    Snapshot 2026-11-06, dated 6 November 2026. #shopify-share

11. 88.4%

    Shopify stores that returned a checkout link

    7,171 of 8,115 completed Shopify store checks. Shopify's agent catalog endpoint is on by default for eligible stores (Shopify's docs, not an ARC measurement).

    Snapshot 2026-11-06, dated 6 November 2026. #shopify-checkout

12. 1.9%

    Stores on other or undetected platforms that returned a checkout link

    49 of 2,583 completed store checks not on Shopify. Every platform other than Shopify, plus stores where no platform was detected.

    Snapshot 2026-11-06, dated 6 November 2026. #other-checkout

13. 56.3%

    Completed checks in United States

    6,024 of 10,698 completed store checks. The directory is not a random sample of the web. The full market split is below.

    Snapshot 2026-11-06, dated 6 November 2026. #top-market

Platform split

Completed checks by detected platform, 6 November 2026. Platform detection is from public signals.

Market split

Completed checks by market, 6 November 2026.

What changed in 2026

These are not ARC measurements. Each one is a company or standards-body statement, linked to its primary source.

• 2026-01-08: Microsoft launched Copilot Checkout in the US, with Shopify merchants enrolled and PayPal and Stripe merchants able to apply. Source: Microsoft Advertising blog.
• 2026-01-11: Google announced the Universal Commerce Protocol (UCP). Checkout in AI Mode runs on UCP for select US merchants that Google approves. Source: Google Merchant Center Help.
• 2026-03-11: Amazon expanded Shop Direct to take product feeds from other retailers' sites, through feed partners such as Feedonomics and Salsify. Source: About Amazon.
• 2026-03-24: OpenAI scaled back standalone in-chat Instant Checkout in ChatGPT, moving to product discovery with checkout on the merchant's own site or app. Source: OpenAI.
• 2026-04-28: Google contributed the Agent Payments Protocol (AP2) to the FIDO Alliance to develop it as a standard. Source: Google blog.
• 2026-07-14: The x402 Foundation, which standardises HTTP 402 payments for agents, began operating under the Linux Foundation. Source: Linux Foundation.

Cite this

How these are counted

One automated HTTP check per store, stopping before payment. Opted-out stores and checks that did not finish are left out of every count. A blank or unknown result is not a no and is not a mark against a store. The full method is on the methodology page.

Frequently asked questions

Agentic commerce statistics: measured, dated, cited

Measured, not forecast. In ARC's snapshot of 6 November 2026 (2026-11-06), 10,698 online stores had a completed automated check. For 67.5% of them (7,220), ARC's check got an agent checkout link back, 1.8% of stores with a readable robots.txt block AI crawlers, and 8,193 have a valid UCP profile. Every number below has its denominator, an anchor to cite, and the open CC-BY-4.0 data behind it.

What share of online stores return an AI agent checkout link?

In ARC's snapshot of 6 November 2026, 67.5%: 7,220 of 10,698 fully tested stores returned an agent checkout link (stores that turned ARC's checker away are left out). A link is not an order. ARC stops before payment.

How many stores block AI crawlers in robots.txt?

188 of 10,281 completed checks with a readable robots.txt (1.8%) block AI crawlers, and 116 of 10,281 (1.1%) block AI answer agents. Snapshot 2026-11-06.

How many stores have a valid UCP profile?

8,193 of 10,698 completed checks (76.6%) in snapshot 2026-11-06. 7,220 of those also returned an agent checkout link.

Are these forecasts?

No. Every ARC figure on this page is counted from the dated snapshot in the open dataset, with its denominator. External facts are marked as such and link to their primary source.

Is this every online store?

No. It is ARC's directory of stores plus stores people asked ARC to check. It leans to Shopify, so platform figures are given separately. A check that did not finish is outside every count and is not a no.

How do I cite these numbers?

Cite ARC (arcreport.ai), snapshot 2026-11-06, with the page URL and the figure's anchor. The data is CC-BY-4.0. The citation file is at /CITATION.cff.

On ARC

• Open dataset →
• Citation file →
• UCP adoption: publish, validate, return a checkout link (10,698-store snapshot) →
• How many online stores support agentic commerce? →
• How many stores support UCP? →
• Methodology →

Cite this: ARC, arcreport.ai, snapshot 2026-11-06 (2026-11-06), CC-BY-4.0. Protocol figures: Snapshot 2026-11-06, 6 November 2026. Index browser figures: Snapshot 2026-11-06, 6 November 2026. Platform figures: Snapshot 2026-11-06, 6 November 2026.

Every figure is labelled Measured, Inferred or Not measured. Not measured is not zero. A missing result is not a mark against a store. ARC stops before payment.