1. What Is Agentic Shopping
Agentic Shopping is used to identify traffic coming from AI assistants and AI search entry points, such as ChatGPT, Claude, Perplexity, Gemini, and similar platforms.
When a shopper first searches, asks questions, compares products, or gets recommendations inside one of these AI products and then clicks through to your store or landing page, Attribuly will try to classify that visit as `Agentic Shopping`. This helps you separate AI-driven traffic from organic search, paid ads, social, email, and other channels.
2. What Problem It Solves
With Agentic Shopping enabled, you can more clearly understand:
- which visits came from AI assistants or AI search
- how much browsing, add-to-cart, checkout, and purchase activity came from AI traffic
- which AI platforms are sending higher-quality visitors
- whether AI-driven traffic is worth more content, SEO, or campaign investment
This is especially useful if you are already doing things like:
- building AI visibility around brand terms, product terms, or comparison terms
- creating landing pages or content designed for ChatGPT, Perplexity, Gemini, and similar tools
- evaluating whether AI recommendation traffic contributes to revenue
3. When You Should Use It
Agentic Shopping is a good fit when:
- shoppers discover your brand, products, or content inside an AI assistant and then click to your site
- traffic comes from AI search, AI shopping, AI recommendation, or conversational discovery flows
- you want AI traffic reported separately instead of being mixed into organic or referral traffic
It should generally not be used for:
- traditional paid advertising clicks
- email marketing clicks
- social post or influencer link clicks
- standard search engine traffic without any AI source signal
4. How The System Identifies Agentic Shopping
The system mainly looks for these signals:
- the visit referrer is a supported AI platform
- the link contains `utm_platform`
- the link contains `utm_source`
If one of these signals clearly matches a supported AI platform, the visit can be classified as Agentic Shopping.
Common supported platforms include:
- ChatGPT
- Claude
- Perplexity
- Gemini
- Copilot
- Meta AI
- Grok
- Poe
- Mistral
- DeepSeek
- Kimi
- Doubao
- Tongyi
- Yuanbao
- Zhipu
This list can be expanded over time.
5. Recommended Usage
For more reliable reporting and clearer attribution, the best practice is to add full UTM parameters to every link you want classified as Agentic Shopping.
Recommended format:
```text
utm_source=chatgpt
utm_medium=agentic_shopping
utm_campaign=spring-launch
utm_platform=chatgpt
```
You can also set different values for different AI platforms:
```text
utm_source=perplexity
utm_medium=agentic_shopping
utm_campaign=category-comparison
utm_platform=perplexity
```
Suggested parameter conventions:
- `utm_source`: the specific AI platform, such as `chatgpt`, `claude`, or `gemini`
- `utm_medium`: use `agentic_shopping` consistently
- `utm_campaign`: the campaign name, landing page theme, or content initiative name
- `utm_platform`: use the platform name again for cleaner reporting splits
6. Recommended Link Templates
If you want one consistent standard for users, content teams, SEO teams, partners, or internal operators, you can use templates like these.
Template 1: ChatGPT Traffic
```text
```
Template 2: Perplexity Traffic
```text
```
Template 3: Gemini Traffic
```text
```
7. How To Confirm It Is Working
You can validate the setup with this quick process:
1. Open a test link that includes the parameters.
2. Check the visit in Attribuly real-time events or the related reporting view.
3. Confirm the channel is shown as `Agentic Shopping`.
4. Confirm `source`, `medium`, `campaign`, and `platform` were written as expected.
If you rely only on the referrer and do not pass `utm_*` parameters, the visit may still be recognized as Agentic Shopping, but your detailed reporting fields may be incomplete.
So the most reliable approach is still to pass the full parameter set explicitly.
8. Common Examples
Example 1: A Visit From ChatGPT
If a user clicks to your site directly from ChatGPT, that visit should typically be recognized as Agentic Shopping.
Example 2: The Referrer Is Google, But The Link Explicitly Marks Gemini
If the final referrer is Google but the link includes `utm_platform=gemini`, the system can still treat it as Agentic Shopping instead of standard organic search.
Example 3: AI Source Exists, But No UTM Parameters Are Added
If the source is clearly an AI platform, the system may still classify the visit as Agentic Shopping. However, if you do not pass `utm_source`, `utm_medium`, `utm_campaign`, and `utm_platform`, your reporting breakdown will be much less detailed.
9. Frequently Asked Questions
Why did AI traffic not get classified as Agentic Shopping?
Common reasons include:
- the link did not include clear parameters
- the traffic passed through an intermediate redirect and the AI referrer was lost
- the visit was classified earlier as ad, email, social, or affiliate traffic
- the platform is not yet on the current support list
Why is the visit classified as Agentic Shopping, but source or campaign is blank?
Because the system may identify the visit from the source page itself, but detailed fields such as source, medium, campaign, and platform usually still need to be passed in through `utm_*` parameters.
How should I separate different AI platforms?
The simplest method is to give each platform its own `utm_source` and `utm_platform`. For example:
- ChatGPT uses `chatgpt`
- Claude uses `claude`
- Perplexity uses `perplexity`
- Gemini uses `gemini`
What naming convention should we standardize internally?
We recommend:
- `utm_medium=agentic_shopping`
- `utm_source` uses the platform standard name
- `utm_platform` matches `utm_source`
- `utm_campaign` follows a consistent internal campaign naming rule