Agent Tools let AI agents and workflows interact with your organization's business systems.
Instead of only answering questions from knowledge, an AI assistant can use tools to retrieve live information, complete actions, generate documents, and start workflows. For example, a tool can check an employee's leave balance, update an address, create a case, call Workday, or query a custom API.
Experience Designers create and manage tools in Applaud. The same tools can then be used across AI conversations, Case Management, workflows, and supported external AI clients.
Why use Agent Tools?
Some employee requests need more than information from a knowledge base.
For example, an employee might ask:
- "How much annual leave do I have left?"
- "Update my home address."
- "Raise a case with HR."
- "Generate my employment letter."
- "Start my onboarding process."
Agent Tools enable an AI assistant to perform these actions by connecting to the systems that hold the required data or provide the required services.
Tools are reusable, so you can configure an integration once and use it across different Applaud experiences.
How Agent Tools work
Agent Tools are built around four concepts:
| Concept | Description |
| Connector | Defines how Applaud connects to a type of system or service, such as Workday, REST, or MCP. |
| Connection | Contains the information Applaud needs to connect to a specific system, such as the endpoint and authentication details. Multiple tools can use the same connection. |
| Tool | Defines a specific action that an AI agent or workflow can perform. |
| Event | Defines information that can be published or received by a workflow. |
Creating a tool
Creating a tool follows four main stages:
Connect → Discover → Refine → Test
- Connect to the system that provides the capability.
- Discover the available operations or MCP tools.
- Refine the selected operation so an AI agent can use it effectively.
- Test the tool with sample inputs before you activate it.
A tool adds information around the underlying operation, including its name, description, inputs, risk level, and the response returned to the AI agent.
This information helps the AI agent understand when to use the tool, what information it needs, and how to interpret the result.
Where you can use tools
The same tool can support different Applaud experiences.
- AI Assistant: Let employees and HR agents retrieve information and complete actions through conversation.
- Case Management: Use tools while working with cases and case-related AI experiences.
- Workflows: Use tools in Agent and Run Tool nodes to perform actions as part of an automated process.
- External AI clients: Allow supported AI clients to access Applaud through the Model Context Protocol (MCP).
This means you can build an integration once and reuse it across multiple experiences, rather than creating a separate integration for each.
Types of tools
Agent Tools support several types of tools.
| Type | Description |
| API tools | Use operations provided by connected systems, including REST, OData, GraphQL, SOAP, and Applaud APIs. |
| MCP tools | Use capabilities provided by an external MCP server. |
| Interactive API tools | Collect and validate information across multiple conversation turns before completing an action. |
| Document generation tools | Generate documents from templates using Applaud's document generation capabilities. |
| Start workflow tools | Start a user-initiated workflow from an AI conversation or another supported experience. |
| Visual tools | Present data as a visual, such as a chart, carousel, map, or other supported visuals. |
Tool ownership
Tools can be managed in different ways.
| Type | Description |
| Standard | Provided and managed by Applaud. Standard tools are available to your tenant but can't be edited or deleted. |
| Custom | Created and managed in your tenant. You can configure, test, activate, and delete custom tools when they're not in use. |
| Workflow-managed | Created and managed by a workflow. Change the tool's behavior in the workflow rather than as a separate tool. |
What users experience
Users don't interact with the Agent Tools administration experience.
When a tool runs during a conversation, the user sees an activity message while the action is in progress. They don't see connection credentials or other configuration details.
Some tools require the user's own account to complete an action. In these cases, the user is prompted to connect their account before the tool can continue.
Benefits
Agent Tools help you:
- Connect AI to live systems. Let AI assistants retrieve current information and perform actions.
- Reuse integrations. Use the same tool across conversations, cases, workflows, and supported external AI clients.
- Control AI actions. Define what a tool does, what information it needs, and how it behaves.
- Test before activation. Validate tools with sample inputs before making them available.
- Build richer experiences. Combine tools with knowledge, conversations, and workflows to complete tasks from start to finish.
Who uses Agent Tools?
| Audience | How they use Agent Tools |
| Experience Designers and tenant administrators | Create connections, configure tools and events, test tools, and activate them. |
| Workflow designers | Add tools to workflows and work with workflow-managed tools. |
| Employees | Use tools through AI conversations without accessing the Agent Tools administration experience. |
| HR and case agents | Use tools through AI experiences in Case Management. |
| External AI users | Access Applaud through supported inbound MCP integrations. |