# Where is AI search traffic coming from? (week of 2026-09-24)

A weekly, measured breakdown of which AI engines and agents are visiting sites, and what it changes about how you publish.

Canonical: https://robauto.ai/blog/ai-search-insights-2026-09-24

Robauto Team · Published Sep 24, 2026 · Last updated Sep 24, 2026

A weekly, measured breakdown of which AI engines and agents are visiting sites, and what it changes about how you publish.

## Where is AI search traffic actually coming from right now?

**97,700 AI agent visits in the last seven days** across the Robauto network, split by the engine that sent them. This is measured traffic from the [Robauto pixel](https://robauto.ai/pixel) — not a survey, not a forecast.

AI agent visits by engine, 7 days ending 2026-09-24. Source: Robauto pixel.Engine / sourceVisits (7d)Share of AI visitsreferral36,04337%direct17,84918%meta10,73311%ahrefsbot8,5499%youtube7,2837%yandex_render6,0646%semrushbot3,2013%google_organic2,8973%crawler2,5723%copilot2,5093%

## Which agents are doing the crawling?

The named agents hitting the most pages this week:

Most active named agents, 7 days ending 2026-09-24. Source: Robauto pixel.AgentVisits (7d)Sites touchedMeta External Agent10,7491,047AhrefsBot8,549323Yandex Render Bot6,064299SemrushBot3,204933Bingbot (Copilot)2,507203

## 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`. Any agent can call it with no key to get the current numbers instead of the snapshot in this post, with a source page behind every value.

- `readiness_index` — current network totals and the average readiness score
- `get_brand_profile` — the grounded record for any brand, with a source page per value
- `get_brand_reachability` — which published route an agent should use to reach a brand
- `agent_scan` — a readiness scan of any domain on demand

Setup and the full tool list are in the [Robauto MCP directory](https://robauto.ai/mcp-directory) and the [MCP server docs](https://robauto.ai/mcp-server).

## What this means for your site

Two things follow from the split above:

- **The engines sending traffic are not the ones most teams optimise for.** If your content only works for one assistant, the rest of that table is traffic you never see.

- **Agent visits arrive with no cookie, no session and often no rendering.** Whatever a crawler can read in raw HTML is your entire pitch.

Measure yours with a free [Agent Readiness Scan](https://robauto.ai/scan), then read [AI Search (AEO/GEO)](https://robauto.ai/learn/ai-search-aeo) for the fixes in order of impact.

The live version of this data, with JSON and CSV downloads under a CC BY 4.0 licence, is on [AI Search Data](https://robauto.ai/ai-search-data). Per-agent detail lives on the [Agent Leaderboard](https://robauto.ai/agent-leaderboard).

[All blog posts](/blog)

---

(c) 2026 Robauto, Inc. — support@robauto.ai
Machine surfaces: https://robauto.ai/llms.txt · https://robauto.ai/llms-full.txt · https://robauto.ai/.well-known/api-catalog
