Skip to main content

Creating and Using Custom Agents

ElixirData allows you to create, configure, and personalize custom AI Agents tailored to your specific business or operational needs.
This guide walks you through creating a new agent, defining its instructions, interacting with it, and updating its behavior over time.

Creating a New Agent​

In the sidebar, hover over the Agents tab. click the Create Agent

A form or modal window will appear where you can define the details of your new agent.

access New Agent Modal

Fill in the following fields:​

FieldDescription
Agent NameEnter a name for your agent (e.g., Agent Booking).
DescriptionAdd a short summary describing what this agent is designed to do (optional but helpful).
InstructionsProvide detailed guidance or behavioral context — this defines how the agent should act or respond to user queries.

For example:

“You are a travel booking assistant. Help users find the best and most affordable hotel options in their selected location.”

Create New Agent Modal

Once you’ve entered all the necessary details, click Create Agent.

Viewing the Created Agent​

After creation, your new agent will automatically appear in the Agents List, alongside other agents.

You’ll see:

  • Agent Name
  • Description
  • Context / Instructions Summary

Agent Context View

Interacting with Your Agent​

Once your custom agent is set up, you can begin chatting directly from its dedicated page.

  1. Open your agent (e.g., Agent Booking).
  2. Type a query or command in the input box.
  3. Press Enter or click Send to start the conversation.

For example:

“Can you provide me cheaper hotels for Goa location?”

The agent processes your request using the defined instructions and generates a relevant response.

Interacting with Agent

Continuing the Conversation​

ElixirData agents maintain contextual memory during each session, allowing natural, multi-turn interactions.

You can continue asking follow-up questions in the same thread to refine your query further.

Example follow-up:

“Show only beachside hotels under ₹3,000 per night.”

The agent remembers:

  • Your previous questions
  • The context and instructions
  • And any attached data sources

This makes conversations dynamic, natural, and context-aware.

Agent Follow-Up Query Example

Viewing Conversation ID and Exporting Responses​

Every conversation with an agent is assigned a unique Conversation ID — allowing you to easily reference or share that specific interaction later.

  1. Inside the conversation thread, look at the top-right corner of the response panel.
  2. You’ll see a unique Conversation ID — click Copy ID to copy it to your clipboard.
  3. You can use this ID to share, revisit, or track the conversation.

Conversation ID View

Exporting Responses from the Thread​

In the same conversation thread, you can directly export the response in multiple formats for reporting or documentation purposes.

  1. Click the Export (⬇️) icon next to the share button.
  2. The exported file automatically includes the Conversation ID, timestamp, and agent details for reference.

Editing and Updating Agent Instructions​

Over time, you might want to adjust your agent’s behavior or update its knowledge.

To do so:

  1. Go to your Agent’s Context section.
  2. Click Edit Instructions.
  3. Update or refine the existing instructions with new details or improved prompts.
  4. Click Save Changes — updates take effect immediately for all future interactions.

Edit Agent Instructions

Best Practices for Creating Custom Agents​

  • Use clear and detailed instructions to shape how your agent responds.
  • Regularly refine your agent’s context for improved accuracy.
  • Test your agent’s responses using different queries before deploying it widely.
  • Keep your agent’s purpose focused and domain-specific to maximize relevance.
💡 Quick Tip
  • Experiment with various instruction styles to fine-tune agent tone and depth.
  • Include example queries within your instructions for better guided behavior.
  • Use short, descriptive agent names to make them easily discoverable.

✨ Summary:
Creating and using custom agents in ElixirData is simple yet powerful.
You can define, instruct, and refine AI agents that align perfectly with your workflows.
With contextual learning and real-time adaptability, custom agents empower teams to automate domain-specific tasks effortlessly.