You can begin automating customer service in just minutes with AI agents. Start with an AI agent that answers customer questions using content from trusted knowledge sources. From there, you can move on to more advanced features such as goal-oriented generative procedures, scripted dialogues, authorized actions, and API integrations with systems you already use. Through it all, advanced analytics are available to help you measure your automation success.
This article gives you an overview of how to start using the most important AI agent functionality.
This article contains the following topics:
Step 1: Optimize your help center content for AI agents
During conversations with customers, AI agents can use content from your trusted knowledge sources, like your Zendesk help center or an external knowledge base, to generate responses to customer questions. The better your content is, the better your AI agents’ responses will be.
If you're just getting started with your help center, first review tickets and other resources to find issues to populate it. See Best practices for finding customer issues to start your knowledge base.
If you already have a help center, consider optimizing your content. See Best practices: Preparing your help center for generative AI.
Step 2: Configure the channels you want your AI agent to work on
AI agents work on messaging, email, API, web form, and voice (EAP) channels to deliver answers to customers. But before they can do that, you need to set up the channels you want your AI agent to work on.
If you haven't already configured these channels, see the following resources:
- Messaging: Getting started with messaging for your website, help center, and mobile apps
- Email: A complete guide to understanding email in Zendesk
- API: Managing API token access to the Zendesk API
- Web form: Understanding and optimizing ticket forms
- Voice: Activating the voice channel and configuring general settings
Step 3: Create your AI agent
With your knowledge sources optimized and your channels configured, it’s time to create your AI agent. As part of the creation process, you’ll:
- Select the brand and knowledge sources the AI agent should use.
- Personalize the AI agent, including defining a name, tone of voice, and supported languages.
- Customize the AI agent’s default system replies.
- Activate the AI agent on the channels it should be able to interact with customers on. (Optional)
See Creating an AI agent to automatically resolve customer issues.
Step 4: Configure your AI agent to power more complex automation (Optional)
Depending on your automation goals, you may want to configure additional settings for your AI agent after you create it. Performing the creation flow described in the previous section results in an AI agent that’s capable of responding to your customers with AI-generated responses based on the knowledge sources you connect to it. However, with additional configuration, you can power more complex workflows.
For example, you can define specific conversation topics and configure how the AI agent should handle conversations about those topics. You can also create actions that the AI agent is authorized to perform, integrations with other third-party systems you use, and more.
The following configuration steps let you power more complex automation workflows:
Step 4.1: Create use cases to identify customer requests
To allow the AI agent to help customers most effectively, you need to create use cases. Use cases are the mechanism by which AI agents understand what a customer is asking about and connect them with the right procedure or dialogue. Think of use cases as the topics your customers need help with, such as order returns or refund requests.
If you don’t create any use cases, the AI agent forms responses using only the content of your connected knowledge sources, not procedures or dialogues.
See Creating use cases for AI agents to identify what customers are asking about.
Step 4.2: Create generative procedures or dialogues to control responses
After you create use cases, you need to define the flow of the conversation that the AI agent should have with a customer when that use case, or topic, comes up. Conversation flows are determined by either a generative procedure, which is more flexible, or a dialogue, which is more structured.
Generative procedures determine the general flow that an AI agent should follow during a conversation with a customer. Procedures allow conversation flows to adapt to a customer’s responses during a conversation with an AI agent, while still aligning with your business policies.
See Creating generative procedures to shape flexible conversation flows for AI agents.
Alternatively, you can create dialogues that allow the AI agent to respond to customer requests according to scripted conversation flows. Dialogues use defined, branching logic to determine an AI agent's responses and actions during a conversation with a customer.
See Creating dialogues to define scripted conversation flows for AI agents.
For each use case you create, you can decide whether it should trigger a more flexible generative procedure, or else a more prescriptive dialogue. Procedures require less setup and maintenance, but offer less direct control over very fine details. Dialogues offer a lot of control, but require more setup and maintenance.
See Configuring whether a use case triggers a generative procedure or a dialogue.
Step 4.3: Create actions, entities, and API integrations to increase automation (Optional)
With your use cases created and conversation flows defined—in addition to your connected knowledge sources—your AI agent is capable of resolving a wide range of customer requests. However, you can make your AI agent even more effective creating:
- Actions that allow the AI agent to perform actions based on the details of the session or your customer relationship management (CRM) system. See Creating and adding actions for advanced AI agents.
- Entities that hold pieces of information in customer messages that have specific meaning, such as the user’s email address. See Creating entities in conversation flows for advanced AI agents.
- API integrations that leverage information from other third-party systems you use during your workflows. To create these API integrations, you use the integration builder. See About the integration builder for advanced AI agents.
Step 4.4: Configure additional AI agent settings to fine-tune behavior
At any point, you can always continue configuring the AI agent’s settings. See Accessing and viewing settings for AI agents.
Step 5: Activate your AI agent on your chosen channels
At this point, your AI agent is ready to start answering a wide range of customer questions. If you didn’t activate your AI agent during the creation flow, and instead opted to perform more advanced configuration, now is the time to activate your AI agent so that it can begin responding to customers.
Step 6: Monitor your AI agent’s performance
The final and ongoing step in creating an AI agent is to monitor its performance and continue to improve its efficiency. The Reporting dashboard gives you insights into key automation metrics, reports on the value of AI agent outcomes, and helps you make data-driven decisions to improve future efficiency.
See Analyzing AI agent performance with the reporting dashboard.