> ## Documentation Index
> Fetch the complete documentation index at: https://docs.useagentshop.com/llms.txt
> Use this file to discover all available pages before exploring further.

# AI Attribution: Why Your Revenue Numbers Are Incomplete

> AI sessions often arrive as Direct in GA4, undercounting real revenue. Here's why attribution breaks and how agentShop's two-source approach fills the gap.

AI-driven revenue is systematically **undercounted** by standard analytics — and understanding why is key to reading every number in agentShop correctly.

When a shopper arrives from an AI assistant, the referral information is often stripped. Google Analytics only credits AI sessions that arrive with an intact referrer; the rest — roughly 35–70% — get filed under **"Direct."** So GA4 alone shows only a fraction of the revenue AI actually drives.

agentShop addresses this with two measurement sources:

* A **first-party pixel** that tags an AI session at the source, capturing more of your true AI traffic.
* Your own **GA4 data**, used as an independent cross-check.

On the **AI Traffic & Revenue** page, the two are shown side by side with the overlap between them. The agreement validates the pixel's number; the difference shows how much GA4 undercounts. The honest takeaway, stated throughout the product: **treat GA4 figures as a floor, not a ceiling.** agentShop never promises perfect attribution — it shows you both numbers and where they differ.


## Related topics

- [AI Traffic and Revenue: Full-Funnel Store Analytics](/user-guide/screens/ai-traffic-revenue.md)
- [Quick Start: Connect Shopify and Run Your First Brand Audit](/user-guide/quickstart.md)
- [agentShop AI Sales Overview: Your Revenue Command Center](/user-guide/screens/ai-sales-overview.md)
- [agentShop User Guide: Get Chosen by AI Assistants](/user-guide/introduction.md)
- [Screen Reference: Every agentShop Screen at a Glance](/user-guide/screens/overview.md)