Agentic Protocol Adoption Census Methodology
For this report, StoreSteady evaluated a frozen corpus of 50 Shopify storefronts. This page documents selection, denominator, exclusion, redaction, source, and version rules.
Corpus And Window
Selection rule: Deterministic rank-stratified 50-store tranche from the frozen external_top_500 Shopify storefront corpus; no merchant-identifying examples are published.
Freeze timestamp: 2026-05-21T22:00:01.570Z
Evaluation window: 2026-05-26T00:44:00.000Z to 2026-05-26T01:05:00.000Z
Fields collected: ACP catalog and checkout endpoint probes, UCP profile probe, AP2 capability declaration probe, MCP server manifest probe, A2A Agent Card probe, llms.txt and llms-full.txt commerce-reference probe, public Shopify product JSON feed probe, checkout/session manifest or public docs reference probe, protocol-adjacent crawler directive probe.
Denominator Rules
Every published rate has its own evaluable N and Wilson 95% confidence interval.
Access friction, detector non-evaluability, deterministic conflicts, tentative findings, and confirmed findings are kept separate.
No blended denominator is used in this report.
Metric Quality Table
| Metric | Tier | N | Count | CI | Caveat |
|---|---|---|---|---|---|
| core_agentic_protocol_verified | Headline | 50 | 0 | 0% - 7.14% | Deterministic public probes for ACP, UCP, AP2, MCP, and A2A. Absence or non-evaluability is not counted as a transaction-capability finding. |
| any_agentic_readiness_signal | Supporting | 50 | 2 | 1.1% - 13.5% | Readiness signal includes verified core protocol surfaces, llms.txt readiness, public feed/API references, checkout/session references, or crawler directives. |
| protocol_adjacent_crawler_directive | Supporting | 50 | 2 | 1.1% - 13.5% | Counts explicit public crawler directives for agent/search bots; this is adjacent readiness, not ACP/AP2/UCP/MCP/A2A adoption. |
| core_protocol_referenced_not_machine_readable | Supporting | 50 | 1 | 0.35% - 10.5% | A public well-known surface responded but did not satisfy the machine-readable V1 schema contract. |
| protocol_probe_blocked_or_not_evaluable | Limitation | 50 | 49 | 89.5% - 99.7% | Non-circumventing public probes only. Blocked/not-evaluable states are separated from protocol absence and from verified adoption. |
Redaction Policy
- No store names, raw domains, raw URLs, raw crawl text, or raw AI responses are published.
- No exact promo codes, private feed/admin evidence, or merchant-identifying product examples are published.
- Public examples fail closed unless Caleb explicitly approves the specific exposure.
Related Public Benchmarks
This report is not a placement dashboard, market forecast, AEO forecast, or financial impact estimate. It measures evidence quality and evaluability under the StoreSteady public research protocol.
Sources
- Stripe Agentic Commerce Protocol documentation - ACP context for catalog, cart, checkout-session, delegated-payment, delegated-authentication, orders, and webhook readiness surfaces.
- OpenAI Agentic Checkout Spec - Checkout-flow context for merchant integration with agentic checkout surfaces.
- Universal Commerce Protocol - Public UCP profile context for merchant capability and commerce profile discovery.
- Visa Trusted Agent Protocol - Trusted-agent context for agent identity, merchant visibility, and signed commerce interactions.
- Model Context Protocol specification - MCP context for tool manifest, auth, consent, and server-discovery readiness checks.
- Agent2Agent protocol repository - A2A context for agent-card discovery and interoperable agent capability metadata.
Version History
- 2026-summer.public-v1 (2026-05-26): Initial public report from a 1-3 store smoke and rank-stratified 50-store deterministic protocol probe tranche.