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

# Configure AI agent skills and actions

> Set up skills so your BluBash AI agent can add labels, create notes, transfer conversations, and call external systems automatically.

Skills define the actions your AI agent can perform automatically during conversations. Rather than simply replying to messages, an agent with properly configured skills can transfer conversations to the right team, label contacts, log notes for your staff, and even call external APIs — all without manual input.

## What skills enable

* **Task automation** — The agent executes actions without human intervention.
* **Organization** — Labels and notes keep conversations categorized and documented.
* **Intelligent routing** — Automatic transfers send customers to the right team.
* **Integrations** — HTTP requests connect the agent to external systems.
* **Efficiency** — Reduces the manual workload for your team.

## How to access skills

<Steps>
  <Step title="Open AI Agents">
    In the left sidebar, click **AI Agents**.
  </Step>

  <Step title="Select your agent">
    Click on the agent you want to configure.
  </Step>

  <Step title="Open the Skills tab">
    Click the **Skills** tab at the top of the page.
  </Step>

  <Step title="Add a skill">
    Click **Add first skill** or **Add skill**.
  </Step>
</Steps>

## Available skill types

### Add Note

Creates an internal note on the conversation with relevant information.

**When to use it:**

* Logging important details about the customer
* Documenting specific requests
* Leaving observations for the human team

**How to configure:**

<Steps>
  <Step title="Select Add Note">
    Choose **Add Note** from the skill type list.
  </Step>

  <Step title="Name the skill">
    Enter a descriptive name.
    Example: *"Log interest in Premium plan"*
  </Step>

  <Step title="Define when to use it">
    Describe the condition that triggers this skill.
    Example: *"When the customer shows interest in the Premium plan"*
  </Step>

  <Step title="Write the note content">
    Enter what will be recorded in the note.
    Example: *"Customer interested in the Premium plan. Requested more details about features."*
  </Step>

  <Step title="Save">
    Click **Save**.
  </Step>
</Steps>

***

### Add Label

Applies a label to the conversation for organization and filtering.

**When to use it:**

* Categorizing the type of request
* Marking interest in specific products
* Identifying priority level
* Segmenting customers

**How to configure:**

<Steps>
  <Step title="Select Add Label">
    Choose **Add Label** from the skill type list.
  </Step>

  <Step title="Name the skill">
    Enter a descriptive name.
    Example: *"Mark as Qualified Lead"*
  </Step>

  <Step title="Define when to use it">
    Describe the trigger condition.
    Example: *"When the customer provides contact information and shows genuine interest"*
  </Step>

  <Step title="Select the label">
    Choose which label to apply from the available options.
  </Step>

  <Step title="Save">
    Click **Save**.
  </Step>
</Steps>

<Tip>
  Create labels specifically for AI-triggered actions before configuring this skill. Examples: "AI - Qualified Lead", "AI - Technical Support", "AI - Pricing Inquiry".
</Tip>

***

### Remove Label

Removes a specific label from the conversation.

**When to use it:**

* Updating the conversation status
* Removing temporary markers
* Correcting miscategorizations

**How to configure:**

<Steps>
  <Step title="Select Remove Label">
    Choose **Remove Label** from the skill type list.
  </Step>

  <Step title="Name the skill">
    Enter a descriptive name.
  </Step>

  <Step title="Define when to use it">
    Describe the trigger condition.
  </Step>

  <Step title="Select the label to remove">
    Choose which label should be removed.
  </Step>

  <Step title="Save">
    Click **Save**.
  </Step>
</Steps>

***

### Transfer to Team

Transfers the conversation to a specific human team.

**When to use it:**

* Routing to a specialized team
* Escalating to technical support
* Forwarding to sales
* Handing off to a human agent

**How to configure:**

<Steps>
  <Step title="Select Transfer to Team">
    Choose **Transfer to Team** from the skill type list.
  </Step>

  <Step title="Name the skill">
    Enter a descriptive name.
    Example: *"Transfer to sales team"*
  </Step>

  <Step title="Define when to use it">
    Describe the trigger condition.
    Example: *"When the customer asks about pricing or wants to close a deal"*
  </Step>

  <Step title="Select the team">
    Choose which team should receive the transfer.
  </Step>

  <Step title="Add a transfer reason (optional)">
    Explain the reason for the transfer if needed.
  </Step>

  <Step title="Save">
    Click **Save**.
  </Step>
</Steps>

<Info>
  After a transfer to a human team, the AI agent stops responding and the team takes over the conversation.
</Info>

***

### Transfer to AI Agent

Transfers the conversation to a different AI agent.

**When to use it:**

* Routing to a specialized agent
* Moving from a triage agent to a specific service agent
* Escalating to a more advanced agent

**How to configure:**

<Steps>
  <Step title="Select Transfer to AI Agent">
    Choose **Transfer to AI Agent** from the skill type list.
  </Step>

  <Step title="Name the skill">
    Enter a descriptive name.
  </Step>

  <Step title="Define when to use it">
    Describe the trigger condition.
  </Step>

  <Step title="Select the target agent">
    Choose which AI agent should receive the transfer.
  </Step>

  <Step title="Save">
    Click **Save**.
  </Step>
</Steps>

***

### HTTP Request

Integrates with external systems through API calls.

**When to use it:**

* Querying information from an external system
* Creating records in a CRM
* Sending notifications
* Integrating with third-party platforms

**How to configure:**

<Steps>
  <Step title="Select HTTP Request">
    Choose **HTTP Request** from the skill type list.
  </Step>

  <Step title="Name the skill">
    Enter a descriptive name.
  </Step>

  <Step title="Define when to use it">
    Describe the trigger condition.
  </Step>

  <Step title="Enter the URL">
    Provide the API endpoint address.
  </Step>

  <Step title="Choose the method">
    Select GET, POST, PUT, or DELETE.
  </Step>

  <Step title="Add headers (optional)">
    Include any required request headers.
  </Step>

  <Step title="Add a body (optional)">
    For POST or PUT requests, enter the request body.
  </Step>

  <Step title="Save">
    Click **Save**.
  </Step>
</Steps>

<Warning>
  This skill requires technical knowledge of APIs. Make sure you have the credentials and documentation for the API you want to integrate before configuring it.
</Warning>

## Best practices

1. **Be specific with conditions** — Clearly define when each skill should be triggered. Use concrete examples and avoid ambiguous descriptions.
2. **Use descriptive names** — Names that explain what the skill does make future maintenance much easier.
3. **Test before activating** — Use the agent test feature to confirm skills are triggered correctly and adjust conditions if needed.
4. **Combine skills** — An agent can have multiple skills. They can run in sequence — for example, adding a label and then transferring to a team.
5. **Keep skills organized** — Periodically review and document the purpose of each skill.

## Managing skills

**Edit a skill:**

1. Find the skill in the list.
2. Click the edit icon.
3. Make your changes.
4. Click **Save**.

**Delete a skill:**

1. Find the skill in the list.
2. Click the trash icon.
3. Confirm the deletion.

**Enable or disable a skill:** Use the toggle next to the skill to temporarily activate or deactivate it without deleting it.

## Important notes

* Skills execute automatically whenever their trigger conditions are met.
* An agent can have as many skills as you need.
* Skills are executed in the order they were created.
* You can test skills using the agent test feature.
* Misconfigured skills can cause unexpected behavior — always test before activating the agent.