Setting up AI Agents feature via OpenAI

Table of contents

• Integrating OpenAI as an AI provider (separate article)
• Step 1 — Create an AI agent
• Step 2 — Trigger the AI agent from an automated rule
• AI Agents feature overview (separate article)

The AI Agents feature lets you create reusable AI-powered agents and trigger them on tickets via automatic rules. This article walks you through setting up an AI agent that uses OpenAI as its AI provider. For a general introduction to the feature, see the AI Agents feature overview.

Note: Before you begin, your LiveAgent account must be connected to OpenAI. If you have not done so yet, follow the Integrating OpenAI as an AI provider guide first.

Step 1 — Create an AI agent

An AI agent is the central configuration object. It tells LiveAgent which AI provider and model to call, how the AI should process each ticket, and which actions it is allowed to perform. You can create as many agents as you need — one per use case is a common pattern (e.g., one for auto-tagging, one for auto-resolving common FAQ tickets).

1. Navigate to Configuration > AI > AI Agents and click Create AI agent.
2. Enter a descriptive Name (e.g., Auto-classify inbound email). The name appears in rule actions and in the run log, so choose something that clearly identifies the agent's purpose.
3. Under Provider, select the OpenAI integration you set up earlier.
4. Under Model, select the OpenAI model that should process tickets for this AI agent.
5. In the Instructions field, enter the instructions that guide how the AI agent handles tickets. This acts as the system prompt for the model — describe the task, the decision criteria, and how the agent should use the available tools. The instructions can be up to 5,000 characters long.
6. Under Allowed tools, select which actions the AI is permitted to take on the ticket. See the MCP Tools Reference for the full list of available tools.
7. Click Save.

Tip: Grant the AI only the tools it actually needs. For example, an agent whose sole purpose is to tag tickets does not need the Resolve ticket or Assign ticket tools. Keeping the tool list tight reduces the risk of unintended actions and helps conserve your credit budget.

Note: Unlike a FlowHunt-based agent, an OpenAI agent does not require any manual MCP connection. LiveAgent connects the model to its built-in MCP server automatically for each run, so you can ignore the MCP Server Integration and MCP Access Tokens sections on the agent — they are only needed for the FlowHunt integration.

Step 2 — Trigger the AI agent from an automated rule

1. Navigate to Configuration > Automation > Rules (or Time Rules for time-based triggers) and click Create.
2. Set a Trigger for the rule — for example, Ticket is created, Message is received, or Tag is added to ticket.
3. Add any Conditions you need (e.g., only tickets in a specific department, or tickets with a particular subject keyword).
4. In the Actions section, click Add action and select Run AI agent.
5. In the action parameters:
   • AI agent — Choose the agent you created in Step 1.
   • Budget — Select which AI budget this run's tool calls should draw from. See the AI Budgets & Credits article for how budgets work.
6. Click Save rule.

From this point forward, every time the rule fires, LiveAgent will dispatch the selected AI agent against the matched ticket. The run executes in the background — no further manual steps are needed. To review runs, see Monitoring AI agent runs.