Defining Inputs (Schema)

When a user makes an inquiry, the agent extracts key details from their prompt and maps them into the schema fields you define.

Example user prompt “Invite Sarah as an account manager with email [email protected].”

From this prompt, the agent should capture:

• email = "[email protected]"
• organizationRoleType = "ORG_ACCOUNT_MANAGER"
• name = "Sarah"

The schema is how you define these fields so the agent knows what to collect.

Visual Editor

Add fields step by step using the form.

How to build the Invite User schema:

Click + Add Field.

1. Name: email
   • Type: String
   • Description: “The email address of the user to invite.”
   • Mark as Required.
2. Name: organizationRoleType
   • Type: String
   • Description: “The role to assign. Must be one of ORG_ACCOUNT_MANAGER (Account Manager), ORG_ADMIN (Admin), or ORG_MEMBER (Member).”
   • Add enum values:
     • ORG_ACCOUNT_MANAGER
     • ORG_ADMIN
     • ORG_MEMBER
   • Mark as Required.
3. Name: name
   • Type: String
   • Description: “The full name of the invited user.”
   • Leave as Optional.

Code Editor

Define the same fields directly in JSON.

📘

Note that the description adds a human-friendly role name (Admin, Account Manager, Member) the agent will use in the conversation with the user

[
  {
    "key": "email",
    "description": "The email address of the user to invite.",
    "type": "STRING",
    "enumValues": [],
    "isRequired": true,
    "isMasked": false
  },
  {
    "key": "organizationRoleType",
    "description": "The role to assign to the user. Must be one of: ORG_ACCOUNT_MANAGER (Account Manager), ORG_ADMIN (Admin), or ORG_MEMBER (Member).",
    "type": "STRING",
    "enumValues": ["ORG_ACCOUNT_MANAGER", "ORG_ADMIN", "ORG_MEMBER"],
    "isRequired": true,
    "isMasked": false
  },
  {
    "key": "name",
    "description": "The full name of the invited user.",
    "type": "STRING",
    "enumValues": [],
    "isRequired": false,
    "isMasked": false
  }
]