Agent Visits ≠ Agent-Ready: #1 Scores 24, #4 Scores 100

Robauto Team · Published Sep 22, 2026 · Last updated Sep 22, 2026

Sep 22 Robauto scans: marketontology.com leads agent visits at score 24 while metareva.com hits 100. AI traffic is 21.18%; readiness median stays 18.

On September 22, 2026, Robauto’s Brand Leaderboard and live Agent Scans disagree in a useful way. marketontology.com leads the network with 9,192 agent visits over seven days and scores 24/100. metareva.com sits fourth with 911 visits and scores 100/100 — Signal Basics, Technical Groundwork, Agent Integration, and Agentic Commerce all complete, with a verified Robauto pixel.

That split is the story. Agent visits measure attention. Agent Readiness measures capability. A site can attract crawlers and assistants without publishing the machine surfaces those agents need to act.

Network pulse the morning of Sep 22

AI agent share of page views edged above 21% on the public summary updated at 2026-09-22T15:00:08Z. The Index median across the most-visited cohort remains 18 — blocked domains are excluded from averages because a block measures scanner access, not readiness.

{
  "as_of": "2026-09-22T15:00:08.773+00:00",
  "ai_search_90d": {
    "total_page_views": 1264824,
    "ai_agent_views": 267867,
    "ai_traffic_percentage": 21.18,
    "top_engines": [
      {"engine": "Meta AI", "visits": 82257, "share_percentage": 30.7},
      {"engine": "grok_bot", "visits": 27008, "share_percentage": 10.1},
      {"engine": "baiduspider", "visits": 23977, "share_percentage": 9.0}
    ],
    "source": "https://robauto.ai/ai-search-data"
  },
  "agent_readiness_index": {
    "attempted": 99,
    "scored": 79,
    "blocked": 20,
    "median": 18,
    "mean": 17,
    "min": 2,
    "max": 29,
    "last_run": "2026-09-02T17:55:37.973493+00:00",
    "source": "https://robauto.ai/agent-readiness-index"
  },
  "live_scans_2026_09_22": [
    {"domain": "marketontology.com", "visits_7d": 9192, "rank": 1, "score": 24, "tiers": {"signal": "6/6", "technical": "0/4", "integration": "0/10", "commerce": "0/5"}, "scan": "https://robauto.ai/scan?url=marketontology.com"},
    {"domain": "thearknewspaper.com", "visits_7d": 1992, "rank": 2, "score": 34, "tiers": {"signal": "6/6", "technical": "0/4", "integration": "0/10", "commerce": "0/5"}, "scan": "https://robauto.ai/scan?url=thearknewspaper.com"},
    {"domain": "topleftdesign.com", "visits_7d": 1855, "rank": 3, "score": 30, "tiers": {"signal": "5/6", "technical": "0/4", "integration": "0/10", "commerce": "0/5"}, "scan": "https://robauto.ai/scan?url=topleftdesign.com"},
    {"domain": "metareva.com", "visits_7d": 911, "rank": 4, "score": 100, "tiers": {"signal": "6/6", "technical": "4/4", "integration": "10/10", "commerce": "5/5"}, "scan": "https://robauto.ai/scan?url=metareva.com"},
    {"domain": "frauddeflect.com", "visits_7d": 141, "rank": 11, "score": 32, "tiers": {"signal": "4/6", "technical": "1/4", "integration": "0/10", "commerce": "0/5"}, "scan": "https://robauto.ai/scan?url=frauddeflect.com"},
    {"domain": "shopify.com", "visits_7d": null, "rank": null, "score": 24, "tiers": {"signal": "6/6", "technical": "0/4", "integration": "0/10", "commerce": "0/5"}, "scan": "https://robauto.ai/scan?url=shopify.com"},
    {"domain": "robauto.ai", "visits_7d": null, "rank": null, "score": 98, "tiers": {"signal": "6/6", "technical": "4/4", "integration": "9/10", "commerce": "5/5"}, "scan": "https://robauto.ai/scan?url=robauto.ai"}
  ],
  "leaderboard_source": "https://robauto.ai/leaderboard"
}

Live scans: traffic leaders vs a fully ready peer

We re-ran Robauto agent_scan (25 read-only checks) on September 22, 2026. Four of the top Brand Leaderboard domains by 7-day agent visits, plus frauddeflect.com, shopify.com, and robauto.ai for contrast:

Read the chart left to right. Cyan bars are visits; violet (or green at 100) bars are scores on a 0–100 scale. Visits and readiness do not move together. metareva.com is the constructive counterexample: fewer visits than the top three, full stack published — MCP server card, A2A agent card (“Meta Reva B2B Agent” with 3 skills), WebMCP, DNS-AID, x402, ACP, UCP, AP2, and MPP.

What “24” usually means on a traffic leader

For marketontology.com, thearknewspaper.com, and shopify.com’s corporate homepage, Signal Basics is complete (HTTPS, robots.txt posture, sitemap, server-rendered text, metadata, and usually llms.txt). Everything above that layer — Link headers with agent relations, API catalog, OpenAPI, markdown negotiation, auth docs, OAuth discovery, A2A, MCP, skills, Content Signals, Web Bot Auth, WebMCP, DNS-AID, and commerce rails — fails. Agents can find and read the site. They still lack a declared way to call it.

topleftdesign.com is slightly different: pixel-verified (hence 30 rather than a mid-20s Signal-only score) but missing llms.txt and every higher tier. frauddeflect.com already serves Accept: text/markdown correctly — one Technical Groundwork pass — and still has no Agent Integration or commerce discovery.

Why this matters while AI traffic climbs past 21%

Builders often treat leaderboard rank as a proxy for “agent-ready.” It is not. Rank answers “who did agents hit?” Readiness answers “what can agents do once they arrive?” With Meta AI alone at 30.7% of the classified AI mix on Robauto’s 90-day summary, discovery pressure is real. Capability remains scarce: Index median 18, and today’s leaderboard re-scans still show zero Integration/Commerce on the #1–#3 visit sites.

Industry guidance is converging on the same split. Chrome’s WebMCP docs treat declared tools as the reliable alternative to fragile DOM actuation. Retail coverage of ACP (OpenAI/Stripe) and UCP (Google and partners) frames commerce as a separate discovery and checkout layer — not something llms.txt can replace. Robauto’s scan simply measures whether those surfaces exist on the live site.

A constructive checklist for this week

1. Open robauto.ai/leaderboard and your own Agent Scan. Treat visits and score as two different KPIs.
2. If Signal Basics is incomplete, finish that first — agents that cannot read you cannot use you.
3. If you already have llms.txt and still score in the 20s–30s, pick one Agent Integration win: WebMCP for on-page actions, or an MCP server card for tool attachment.
4. Only open ACP/UCP/x402/AP2 when you intend agents to transact. metareva.com shows the full ladder is shippable on a public domain today — not theoretical.
5. If agents are not welcome, say so in robots.txt and Content-Signals. A low score under an explicit block is policy working, not a failure.

What this means for builders

On Sep 22, Robauto’s network shows AI agents at 21.18% of page views over 90 days, while the Agent Readiness Index median stays 18/100. The Brand Leaderboard’s #1 domain draws 9,192 weekly agent visits at score 24. A fourth-place peer scores 100 with commerce rails live. Visits prove agents showed up. Readiness proves you gave them something to do. Build for both — and measure them separately.

Sources

• Robauto Brand Leaderboard
• Robauto AI Search Data
• Robauto Agent Readiness Index
• Agent Scan — marketontology.com
• Agent Scan — thearknewspaper.com
• Agent Scan — topleftdesign.com
• Agent Scan — metareva.com
• Agent Scan — frauddeflect.com
• Agent Scan — shopify.com
• Agent Scan — robauto.ai
• Chrome WebMCP docs
• llms.txt
• Model Context Protocol
• Agentic Commerce Protocol
• UCP overview
• x402 docs
• ACP vs UCP guide (Hashmeta)

Where can an agent get this data live?

Robauto MCP — Agent Intelligence & Brand Discovery Infrastructure. A public, stateless Streamable HTTP endpoint at https://mcp.robauto.ai/mcp, callable with no key. Setup and the full tool list are in the Robauto MCP directory.

All blog posts