> For the complete documentation index, see [llms.txt](https://docs.fullwhere.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.fullwhere.com/documentation/documentation-en/automations/attributes.md).
# Attributes
{% hint style="success" icon="list-check" %}
**IN BRIEF**
Create custom attributes that let you define your business logic directly in Fullwhere.
These attributes can be linked to your customer interactions, your customers, and their purchases.
Attributes can be manual fields or fields whose value is filled in by AI Completion.
Attributes can be used to create fully customized conditional logic in your Workflows.
{% endhint %}
{% hint style="info" icon="shield-check" %}
**ACCESS RIGHTS**
To modify the Attributes in your Fullwhere workspace, you must be a supervising member with Administrator rights.
{% endhint %}
## Access your attributes
To create or modify your custom attributes, go to Settings > [Attributes](https://app.fullwhere.com/automation/attributes).
The first table lets you access your attributes based on their target, namely:
* Interactions: custom attributes linked to your customer interactions
* Customers: custom attributes linked to your customers
* Purchases: custom attributes linked to your customers' transactions
By clicking on a category, you access the attributes for the chosen category. The attributes of that category are then presented in the following table:
Column
Description
Name
Name you gave your attribute
Type
The type of your attribute (Text, Number, Single Select, Multiple Select)
AI Completion
Indicates whether AI completion is enabled (available only for attributes targeting interactions)
## Create an attribute
To create a custom attribute, click the button `+ Create an attribute` at the top of your screen.
In your attribute configuration menu, you can define the following settings:
### Type
The type of an attribute determines the values that can be set on that attribute. There are 4 different types.
#### Text
An attribute of type **text** can be assigned any free-form value from one entity to another, whether by users in your account or by AI Completion if it is enabled.
Text-type attributes can be useful, for example, for taking notes or creating summaries.
#### Number
An attribute of type **number** can be assigned only a numeric value, whether by users in your account or by AI Completion if it is enabled.
Number-type attributes can be useful, for example, for defining custom scoring systems.
#### Single Select
An attribute of type **single select** can be assigned a single value from a list of predefined possible values, whether by users in your account or by AI Completion if it is enabled.
If the[AI Completion](#ai-completion) is enabled on your attribute, you can [allow AI to generate new values](#autoriser-lia-a-generer-de-nouvelles-valeurs).
Single-select attributes can be useful, for example, for defining a priority level or the team concerned by the interaction.
#### Multiple Select
An attribute of type **multiple select** can be assigned multiple values from a list of predefined possible values, whether by users in your account or by AI Completion if it is enabled.
If the[AI Completion](#ai-completion) is enabled on your attribute, you can [allow AI to generate new values](#autoriser-lia-a-generer-de-nouvelles-valeurs) or not.
Multiple-select attributes can be useful, for example, for detecting the products mentioned by your customers in your interactions.
{% hint style="info" %}
Once your attribute is created, the type of your attribute can no longer be changed.
{% endhint %}
### Concerned entity
Your attribute can concern one of the following 3 entities:
* Interactions: custom attributes linked to your customer interactions
* Customers: custom attributes linked to your customers
* Purchases: custom attributes linked to your customers' purchases
{% hint style="info" %}
Once your attribute is created, this setting can no longer be changed.
{% endhint %}
### Name
The name of your attribute is the value by which you can identify your attribute in your Fullwhere workspace.
If AI Completion is enabled on your attribute, its name will be provided to the LLM responsible for assigning values to your attribute, so in that case, make sure to use an unambiguous name.
### Possible values
This setting is available only for single select and multiple select attributes.
For each item, you must specify:
* its value
* its color
{% hint style="warning" %}
Modifying a possible value changes it everywhere it was assigned.
{% endhint %}
{% hint style="danger" %}
Deleting a possible value causes it to be deleted on all entities (interactions, customers, purchases) to which it was assigned.
**This action is irreversible.**
{% endhint %}
### AI Completion
{% hint style="warning" %}
AI features are available only for attributes targeting customer interactions.
{% endhint %}
You can take advantage of our AI features to automatically enrich your data with AI Completion. From then on, for each interaction centralized in Fullwhere, the values of your attributes can be defined by AI analysis.
#### Directives
The field **directives** lets you define the prompt that will be used to determine the value assigned to your attribute.
In the directives, it is important to clearly define what your attribute represents and how it should be analyzed by the LLM.
For single select and multiple select fields, it is important in the directives to clearly define a description for each of the field's possible values:
* describe what the value represents and in what circumstances it should be selected
* it may be useful to include keywords or phrases that the customer might use
* it may be useful to specify what does not fall under a possible value, i.e. in which cases it should not be selected
For single-select attributes for which AI Completion is enabled, it is recommended to have minimal overlap between the possible values. If the values are too similar, it may be difficult for an LLM to assign the right one.
Before finalizing, ask yourself the following question: “Would a human have trouble choosing between the possible values?”
{% hint style="info" %}
To validate your directives, it may be useful to submit them to an LLM chat like ChatGPT, Gemini, or Claude to help make them as clear as possible.
{% endhint %}
#### Detection conditions
Detection conditions allow you to perform AI detection on your attribute only if the customer interaction meets the conditions you define.
This can be useful for specializing your attributes and not performing AI detection when it is not necessary.
#### Allow AI to generate new values
For single select and multiple select attributes, you can choose whether or not to allow AI to add new values to the list of possible values.
In this case, the new values will be added to your attribute and will be used for future attribute detections.
#### Overwrite values already present
For multiple select fields, you can specify whether the values detected by AI should be added to the values that may already be present or whether they should replace them.
## Delete an attribute
To delete a custom attribute, go to Automations > Attributes, then select your attribute family.
In the list of attributes, click the 3 little dots at the end of the row for the attribute to delete, then in the dropdown menu click Delete.
{% hint style="danger" %}
Deleting an attribute causes all values that had been assigned to your interactions, customers, or purchases to be deleted.
**This action is irreversible.**
{% endhint %}
***
## Use cases
Here are concrete examples of how to use custom attributes with AI Completion.
### 1. Priority level
This attribute automatically sorts incoming messages to handle urgent ones first.
* Name: `Priority`
* Type: `Single Select`
* Possible values: `Low`, `Medium`, `High`
* Allow AI to create new values: `No`
* **Directives**
> "Analyze the customer interaction to determine its level of urgency and importance.
>
> * Choose High if the customer expresses strong anger, reports a blocking technical issue, or requests cancellation.
> * Choose Medium for common usage questions, order follow-up requests, or non-blocking issues.
> * Choose Low for thank-yous, sharing positive feedback, or purely informational messages that do not require an immediate response."
***
### 2. Custom scoring system
This attribute turns textual sentiment into a numerical value to feed your reports.
* Name: `Satisfaction Score (AI)`
* Type: `Number`
* **Directives**
> "Evaluate the overall level of satisfaction expressed by the customer in their message on a scale from 0 to 20.
>
> * Assign a score close to 0 if the customer is extremely dissatisfied or aggressive.
> * Assign a score of 10 if the message is neutral or purely factual.
> * Assign a score close to 20 if the customer expresses strong enthusiasm or warm thanks.
>
> Respond only with the number corresponding to the score."
***
### 3. Detection of mentioned products
Ideal for analyzing which products generate the most conversations without manual input.
* Name: `Mentioned products`
* Type: `Multiple Select`
* Possible values: `Product A`, `Product B`, `Product C`
* Allow AI to create new values: `Yes`
* **Directives**
> "Identify all specific products or services mentioned by the customer in the interaction text.
>
> * If the customer mentions a product from the list of possible values, select it.
> * If the customer mentions a product that is not yet in the list, create a new value corresponding to the detected product name.
> * If no specific product is identified, do not select any value."
## Use an attribute
The value of your attributes can be defined/modified for each customer interaction present in Fullwhere.\
To define/modify the value of an attribute for a customer interaction:
* go to the customer interaction to edit
* in the information panel on the right, go to the Custom Attributes section
* you can define the value to assign for each attribute
Your attributes can also be used to define logical conditions in your [Workflows](/documentation/documentation-en/automations/workflows.md#conditions-portant-sur-linteraction).
## Analyze attributes
You can explore your attributes in Fullwhere Analytics to analyze their values and better manage your activity.
In other words, your custom fields are not just for storing information, they are real management levers for understanding your data and spotting trends.
This can for example help you to:
* identify the most frequent values
* compare priority, satisfaction, or typology levels
* track how your attributes evolve over time
---
# 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://docs.fullwhere.com/documentation/documentation-en/automations/attributes.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.