Configure an Applaud tool to control the information it sends to and receives from an Applaud API.
You configure the tool's Request, Conditions, and Response so the AI assistant has the information it needs without exposing unnecessary API details.
Before you begin
Configure the request
The Request section defines the information the tool sends to the underlying Applaud API.
- Open the Request section.
- Review the available fields.
- Decide which fields the assistant needs to provide.
- Map each required field to an appropriate value.
- Remove any fields the tool doesn't need.
- Save your changes.
Choose a mapping
Use the appropriate mapping for each request field.
| Mapping | Use it when |
| Agent-provided | The assistant needs to provide the value from the conversation. |
| Fixed | The value should always be the same. |
| Context | The value comes from the signed-in user or case context. |
| Set value | The value comes from a template or expression. |
| Unused | The operation doesn't need the field. |
Use context for the current user
Use Context when the operation needs information that's already known about the person using the assistant.
For example, if the tool answers:
"What recognitions have I received?"
map the employee identity to the signed-in user's context rather than asking the person for their employee ID.
This keeps the experience simple and avoids asking people for information Applaud already knows.
For a tool that works with another employee's records, use an appropriate Agent-provided value, such as a name, email address, or ID, if the underlying operation supports it.
Use internal variables
Use Internal variables for values the tool needs but don't need to be visible or provided by the assistant.
For example, an API might require an internal object ID even though the person only needs to provide a name.
Map required fields
Every required field must have a valid mapping before you can save the tool.
Review each field and make sure:
- The value comes from the right source.
- The assistant doesn't need to ask for information that's already available in context.
- Technical values aren't unnecessarily exposed to the person.
- The mapping sends the value in the format expected by the API.
Add conditions
Use Conditions when the tool needs to change its behavior based on information collected during the conversation.
You can use conditions to:
- Show or hide fields.
- Provide default values.
- Collect additional information only when it's needed.
- Call another tool to retrieve information.
- Use context to identify the relevant person.
For example, a guided request might ask for a location first and then show only the options available for that location.
Configure create and update tools
If your tool changes Applaud data, design the interaction so the employee can review what will happen before the action is submitted.
A typical flow is:
Collect → Confirm → Submit
- Collect the required information.
- Validate the values.
- Show a concise summary of the proposed change.
- Ask the person to confirm.
- Submit the request after confirmation.
For example, an address update might collect the new address, show it back to the employee, and submit the change only after the employee confirms it.
Configure the response
The Response section defines the information returned to the assistant after the tool runs.
The API might return many fields, but the assistant usually needs only a small subset to answer the person's request.
For example, a recognition lookup might return:
- Recognition title
- Date
- Sender
- Message
rather than the complete API response.
- Open the Response section.
- Review the fields returned by the operation.
- Keep the fields the assistant needs.
- Remove unnecessary or sensitive information.
- Reshape or enrich the response when needed.
- Save your changes.
Shape the response for the assistant
Use clear, meaningful fields that help the assistant understand the result.
For example, instead of returning a large API payload containing technical metadata, return a concise structure such as:
- recognition_title
- recognition_date
- recognized_by
- message
Configure tools that use other tools
A tool can use another tool when it needs additional information to complete a request.
For example:
- The assistant identifies the employee.
- One tool retrieves available options.
- The person selects an option.
- The main tool submits the request using the selected value.
When you use nested tools, make sure each tool has a clear purpose and returns only the information needed by the next step.
Review your configuration
Before testing the tool, check that:
- All required request fields are mapped.
- User or case information uses Context where appropriate.
- Internal values aren't unnecessarily exposed to the assistant.
- Conditions only collect information when it's needed.
- Create and update actions include an appropriate confirmation step.
- The response contains the information the assistant needs.
- Unnecessary or sensitive fields aren't returned.
- Any nested tool calls have a clear purpose.