> For the complete documentation index, see [llms.txt](https://alhena.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://alhena.gitbook.io/docs/developer-reference/website-sdk/examples/ab-test-guidelines.md).
# A/B Test Guidelines
This example shows how to configure Alhena AI to respond differently based on whether a user is part of a **Variant** or **Control** group, enabling you to run A/B tests on AI behavior.
## What You'll Learn
* Pass A/B test group assignments via user metadata
* Configure guidelines with Trigger/Action format
* Differentiate AI responses based on group membership
## Use Case
A/B testing your AI agent allows you to:
* Test different greeting styles or conversation openers
* Compare conversion rates between response strategies
* Measure the impact of different AI behaviors on user engagement
* Validate changes before rolling out to all users
For example:
* **Variant Group:** Display a greeting message to the user
* **Control Group:** Ask the user for their email address
## Prerequisites
* Alhena SDK installed on your website
* An A/B testing system that assigns users to groups
## Supported Variables
Alhena supports two metadata variables for A/B testing:
| Variable | Purpose |
| ------------------ | ----------------------------------- |
| `ab_control_group` | Identifies Control Group users |
| `ab_variant_group` | Identifies Variant/Test Group users |
## Code Example
### Adding user to Control Group
If the user is in your control group, pass `ab_control_group` in user metadata:
```html
A/B Test - Control Group
Welcome
```
### Adding user to Variant Group
If the user is in your variant group, pass `ab_variant_group`:
```html
A/B Test - Variant Group
Welcome
```
> **Note:** Don't pass both control and variant group for the same user.
## Configuring Guidelines
Navigate to **AI Settings > Guidelines** and create guidelines using the Trigger/Action format to respond differently based on user group.
### Variant Group Guideline
**Trigger:** Metadata in User Information contains "ab\_variant\_group"
**Action:**
```
Reply with a friendly greeting at the beginning of the conversation. Welcome the user warmly before addressing their question.
```
### Control Group Guideline
**Trigger:** Metadata in User Information contains "ab\_control\_group"
**Action:**
```
Ask for the user's email address at the beginning of the conversation before providing assistance.
```
## How It Works
1. **Group assignment**: Your A/B testing system assigns users to either control or variant group
2. **Metadata passed**: The SDK sends the group assignment via `userMetadata` to the AI.
3. **Guideline triggers**: Alhena detects the metadata and activates the matching guideline
4. **Differentiated response**: The AI follows the action defined for that group
## Customization
### Dynamic group assignment
Integrate with your A/B testing platform:
```javascript
// Example with a hypothetical A/B testing library
var userGroup = ABTestingLib.getGroup('ai-greeting-test');
document.gleenConfig = {
company: 'your-company-key',
apiBaseUrl: 'https://app.alhena.ai',
userMetadata: userGroup === 'variant'
? { "ab_variant_group": "1" }
: { "ab_control_group": "1" }
};
```
### Multiple A/B tests
You can run multiple tests by using descriptive metadata values:
```javascript
document.gleenConfig = {
company: 'your-company-key',
apiBaseUrl: 'https://app.alhena.ai',
userMetadata: {
"ab_variant_group": "greeting_test_v2"
}
};
```
Then create guidelines with more specific triggers:
**Trigger:** Metadata in User Information "ab\_variant\_group" equals "greeting\_test\_v2"
## Related
* [Guidelines](/docs/ai-configuration/tuning/guidelines.md) - Full guideline configuration documentation
* [User Metadata](https://github.com/Gleen-AI/gitbook-docs/blob/main/developer-reference/website-sdk/configuration.md#usermetadata) - All metadata options
* [JavaScript API](/docs/developer-reference/website-sdk/javascript-api.md) - Widget control methods
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:
```
GET https://alhena.gitbook.io/docs/developer-reference/website-sdk/examples/ab-test-guidelines.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.