What are Actions and How are They Used?

Think of Actions as giving your eesel AI assistant hands to interact with your connected platforms like Zendesk, Intercom, Shopify, Jira, and your knowledge base. Instead of just providing information, the AI can take action based on the user's request or the conversation's context.

When a user interacts with your eesel AI, the LLM analyzes the conversation. If the user's intent matches a pre-configured instruction linked to a tool, the LLM will:

• Identify the appropriate tool.
• Extract any necessary information (parameters) from the conversation required by the tool.
• "Call" the tool by passing it the extracted parameters.
• Receive a response from the tool (e.g., confirmation of an action, data from a database).
• Use the tool's response to inform its next interaction with the user.

Why Use Actions in eesel AI?

Integrating Actions into your eesel AI configuration transforms it from a purely conversational agent into a proactive assistant. For support and e-commerce teams, this means your AI can:

• Resolve issues faster: Directly close tickets, tag conversations, or fetch order details without human intervention.
• Improve agent efficiency: Automate routine tasks, provide agents with pre-fetched information, or correctly route complex queries.
• Enhance customer experience: Provide immediate actions like looking up order status or creating necessary follow-up tasks (like bug reports).
• Maintain data hygiene: Ensure conversations and tickets are consistently tagged and organized according to your workflows.
• Automate complex workflows: Chain multiple actions together, like fetching customer details, summarizing the interaction for an internal note, and escalating the conversation to the right team.

By strategically configuring your LLM to use the available Actions, you empower it to handle a wider range of tasks autonomously and integrate seamlessly into your existing operational workflows.

Available Actions and Configuration Guidance

Below is a breakdown of the standard Actions available in eesel AI and how you can instruct your LLM to use them. When configuring your LLM's system prompt or instructions, use clear, conditional statements that tell the AI when to use a tool and what information to provide.

Important Note on Prompts: The examples below illustrate the type of instruction you would include in your LLM's configuration prompt to guide its tool usage. Note that context like the current ticket ID or conversation ID is automatically provided to the AI and does not need to be extracted or passed as a parameter unless explicitly listed below.