Creating and customizing agents

Edit online

Creating and customizing an AI agent involves defining its purpose and personality through thoughtful descriptions, selecting the most suitable AI model, and choosing a style that aligns with your brand or use case. You can also configure key elements like the welcome message, starter prompts, and voice settings to ensure the agent engages users effectively from the first interaction. Together, the configured components shape how the agent communicates, responds, and delivers value across conversations.

Creating agents

You can create a new agent in two ways:

Create from a template: Select an agent from the catalog to use as a template. The agent opens in the Agent Builder for customization and deployment.

Create from scratch: If no suitable template is available, create a new agent from scratch directly within the Agent Builder

To create an agent:

  1. From the chat page, click Create new agent.

  2. Select either Create from scratch or Create from template.

Creating an agent from scratch

Use this option when you want to configure every aspect of the agent without predefined constraints:

  1. Click Create from scratch.

  2. Enter the agent's Name and Description. For more information, see Recommendations for agent descriptions.

  3. Click Create.

  4. Click Deploy to make your agent live.

Creating an agent from a template

Use this option to quickly begin agent creation with a ready-made structure that you can easily tailor:

  1. Click Create from template.

  2. Select an agent from the list.

  3. Click Use as template.

  4. Click Deploy to make your agent live.

Organize agents with workspaces on IBM Cloud

  • You can now create workspaces to organize your agents and collaborate on IBM Cloud.
  • Workspaces give your team a private space where you can build, test, and manage agents together while keeping in‑progress work visible only to the people you choose.
  • You can assign members as owners or editors with role‑based permissions, move agents between workspaces while preserving dependencies, decide who can view or modify specific agents, and structure your work by project, team, or business unit.
  • Workspaces also include analytics that are specific to the workspace you are using, giving you focused insight into the activity and performance of the agents in that space.
  • For more information about creating and managing workspaces, see Overview of workspaces.

Expanding your agent's capabilities

Once your agent is created, you can enrich and extend its functionality through the following configuration options:

Define a profile

Provide a clear and specific description of the agent’s purpose. This helps in multi-agent orchestration by enabling accurate selection based on capabilities. Consider the following practices when defining the description of your agent.

Define a style

Choose how the agent interprets and responds to user requests, selects actions, and engages with tools. See Choosing a style for your agent.

Choosing an AI model

Select an AI model from the Agent Builder UI to define your AI agent’s core capabilities. See Choosing an AI model.

Customize the welcome message and starter prompts

Customize the welcome message and starter prompts to guide users when they begin interacting with the agent. See Customizing the welcome message and starter prompts.

Customize the text below the welcome message

You can now modify the default light gray text below the welcome message only through ADK. For more information, see Customize the light gray text.

Configure voice (Preview)

Enable voice interactions by creating a voice configuration, allowing agents to communicate verbally with users. See Configuring voice settings for agents (Preview).

Add knowledge (Optional)

Enhance the agent’s domain expertise by adding contextual knowledge from files or content repositories. See Adding knowledge to agents.

Add tools (Optional)

Integrate tools to enable the agent to perform automated tasks such as retrieving data or sending emails. See Adding tools to agents.

Add collaborator agents (Optional)

Enable multi-agent orchestration by adding collaborator agents that work together to achieve shared goals. See Orchestrating agents.

Add instructions (Optional)

Define behavioral rules and interaction logic to guide the agent’s responses and actions. See Adding instructions.

Connect to channels

Make the agent accessible to your team by connecting it to various interaction channels. See Connecting to channels.

Deploy the agent

Finalize the setup by deploying the agent to make it available in live environments such as chat interfaces. See Deploying agents