Skip to main content
In this article, we’ll discuss how agents function as dedicated modules that integrate effortlessly into a conversation flow. Agents are equipped to handle specific user requests that go beyond the standard capabilities of the assistant. By delegating these complex tasks to agents, your main assistant remains focused on its core interactions, keeping it efficient and responsive. This setup also allows your system to be modular and adaptable, supporting advanced functionalities without adding unnecessary complexity. Once an agent is built, it can be easily incorporated into an execution flow using the Agent action, expanding your assistant’s range of abilities in just a few steps.

​
Topics covered:

​
Steps to Create a New Agent

  • Navigate to AI Pilots from the left side bar.
  • Choose existing pilot or create New
Screen Shot 2025-06-09 at 4.45.08 PM.png
  • Move to “Agents”
  • Click: **“New” **button
  • Configure New Agent in the Popup
Screen Shot 2025-06-10 at 5.23.44 PM.png

​
How to Configure an Agent

You will be need to provide below asked details in order configure your new agent. **Agent Details **
  • Name: Provide the agent’s name.
  • Description (Optional): Add a brief description of the agent’s purpose or role.

​
Tools

Functions: Functions allow the agent to perform tasks such as interacting with APIs or processing data.
  • Add Functions: Select functions from the library or create new ones.
  • LLM Description (Optional): Provide a description for each function to specify when and how it should be used.
  • **Input Variables: **Define LLM descriptions and default values for input variables.
**Integration: **Enable your Agent to work with both Skara’s built-in modules and external third-party apps.

**By tools: **Leverage Skara’s prebuilt tools to make everyday actions easier. Sub-flows**: **Sub-flows enable the agent to incorporate other flows (optional)-
  • Add Sub-flows: Include synchronous sub-flows from Smart Flows.
  • LLM Descriptions: Provide descriptions for each sub-flow.
  • Configure Variables: Set variables as needed for the sub-flow.
**API Call: **Empower your Agent to access or transmit data to external services through APIs.

​
Paths

Paths define the conversational routes the agent follows based on different outputs or decisions.
  • Add Output Ports (Optional): Define output ports for return values.
  • Define Paths: Assign names, conditions, and variables (with LLM descriptions) for each path.
  • Avoid Duplicate Path Names: Ensure each path has a unique name.

​
Knowledge Base Access

The Knowledge Base (KB) provides additional data the agent can use to provide informed responses.
  • Enable Access (Optional): Allow the agent to access the KB if needed.
  • LLM Description: Describe when and how to use the KB.
HA6.png Prompt Settings
  • Temperature: Use this to adjust the creativity level of AI-generated responses. Use the slider to set the value between **0 **(least creative) and **1 **(most creative).
Screen Shot 2025-06-10 at 5.28.13 PM.png

​
Editing an Agent

  • Navigate to AI Pilots.
  • Choose Your Pilot: Select the relevant Pilot containing the agent you want to edit.
  • Move to Agents: Access the Agents section under your selected Pilot.
  • Edit an Agent: Hover over the agent you want to edit and then click actions.
  • Make Changes: Edit the the agent and then hit “Update”
Screen Shot 2025-06-10 at 5.29.07 PM.png