See AI agent traffic
on your store.
See which AI agents visit, where they drop off, and which fixes may improve conversion. Edge worker and log ingestion give the strongest coverage; beacon mode is a browser-like fallback.
We do not invent revenue.
Lens starts with what it can observe: agent visits, verified signatures, protocol health, journey depth, paths, and products. When order data is connected, it ties those sessions to carts and orders. Until then, dollar figures are labeled estimates.
2,847 demo agent events. 780 reconstructed sessions in the journey sample. The gap between cart and checkout handoff is where the issue appears.
Checkout handoffs are the main drop-off
ChatGPT and Claude sessions reach late-funnel pages, then stall at checkout handoff. Lens shows the path before order data exists.
View estimateProduct data is incomplete for agents
Claude and ChatGPT already show high-intent product reads. Missing structured fields make those sessions harder to route.
Label estimateGoogle Shopping connection slowed to 2.3 seconds
Intermittent UCP errors degrade discovery for Gemini and Perplexity. Lens separates the protocol issue from traffic volume.
Inspect healthExplore the full dashboard.
Traffic charts, agent quality, journey funnels, labeled attribution estimates, session replay, and 9 AI-generated insights. All using representative ecommerce data.
Request-level agent behavior when coverage is installed
Traffic patterns, capture health, compatibility signals, and attribution across the storefront surfaces Lens can observe.
- TrafficWhich agents visit — ChatGPT, Gemini, Claude, Copilot — and how they interact with your store
- JourneysSession timelines showing where agents land, browse, and drop off
- AttributionAgent-attributed carts and orders when order events are connected; estimates are labeled until then
- InsightsPriority fixes with estimated impact when enough traffic evidence exists
- AlertsScore changes, compatibility issues, agent behavior anomalies
- PortfolioMulti-store monitoring — compare stores, prioritize interventions
Start with traffic quality. Attribute revenue only when the data exists.
Lens shows agent traffic, verification quality, and deepest funnel stage now. When cart and order events are connected, the same sessions can be tied to orders.
The demo estimate is benchmark-based and reconciles to the five fix rows below. With order events connected, Lens shows measured agent-attributed carts and orders instead.
Lens is included in every plan.
Deterministic traffic view. Up to 30 days until 500 agent visits, then 24-hour window.
30-day history. AI insights. Order attribution when connected.
90-day history. 100-store portfolio. White-label.
Works from CLI, API, and MCP
AI agents discover and invoke tools programmatically. Developers get the same access from the command line.
# Get Lens statuscurl https://agenticcom.ai/api/v1/lens/sites \ -H "Authorization: Bearer col_..." # Get agent traffic for a sitecurl https://agenticcom.ai/api/v1/lens/events \ -H "Authorization: Bearer col_..." \ -d '{"site": "my-store.com", "period": "7d"}' # Returns: agent_visits, sessions,# verified_share, attribution_status,# top_agents, drop_off_pointsInstall: npx -y @getcolter/cli · MCP: npx -y @getcolter/cli mcp --admin-tools
Common questions
How does Lens detect AI agents?
Lens classifies request-level traffic from the edge worker or server logs using user agent patterns, signed-agent headers, protocol attempts, and request context. The JavaScript beacon is a partial fallback for agents that execute storefront JavaScript.
Does collection slow down my store?
No. The edge worker records metadata asynchronously and forwards the original request to your origin. Beacon fallback is lightweight and asynchronous when used.
What’s included in the free tier?
Free Lens gives one site a deterministic traffic view. You get up to 30 days of history until 500 agent visits, then a 24-hour window. LLM insights, exports, journey analytics, and external alerts are paid features.
Can I see which products agents look at?
Yes. Lens tracks agent journeys at the page level — which products they browse, where they drop off, and which ones lead to cart or checkout when order data is connected.
How does revenue attribution work?
Lens traces agent sessions through discovery, browse, cart, and checkout stages. Revenue attribution starts when cart and order events are connected; before then, any dollar impact is labeled as an estimate.
Start with a free scan. Lens comes next.
Run Check (free), deploy Fix, then activate Lens to monitor real agent traffic.
Scan your store