Prompts

FrontDeskOS provides pre-built prompt templates that guide AI agents through common front desk workflows. Prompts combine instructions, context, and tool suggestions into reusable templates.

What Are Prompts?

MCP prompts are templated instructions that an AI agent can use to handle specific scenarios. Each prompt includes a description, required arguments, and a formatted message that combines the workspace context with the scenario at hand.

handle_incoming_call

The primary prompt for handling incoming phone calls. It instructs the AI to greet the caller, identify their needs, and take the appropriate action (schedule, transfer, take message).

Parameter	Type	Description
`caller_phone`required	`string`	The incoming caller's phone number.
`caller_name`	`string`	Caller name if known from caller ID.
`is_returning`	`boolean`	Whether this is a known/returning caller.

Using the prompttypescript

const prompt = await mcp.getPrompt("handle_incoming_call", {
  caller_phone: "+1-555-0123",
  caller_name: "John Smith",
  is_returning: true
});

// The prompt returns a structured message like:
// "You are the AI receptionist for Riverside Family Dental.
//  A returning patient, John Smith (+1-555-0123), is calling.
//  Their last visit was on Dec 15, 2024 for a dental cleaning.
//  Greet them warmly, ask how you can help, and use the
//  appropriate tools to assist them."

daily_briefing

Generates a morning briefing summary for staff including today's appointments, pending callbacks, and key metrics from the previous day.

Parameter	Type	Description
`date`	`string`	Date for the briefing (defaults to today).
`staff_id`	`string`	Generate briefing for a specific staff member.

lead_followup

Generates a follow-up plan for a lead including suggested talking points, relevant history, and recommended next steps.

Parameter	Type	Description
`lead_id`required	`string`	The lead to generate follow-up for.

after_hours_handler

Handles calls received outside business hours. Informs callers of office hours, offers to take a message or schedule a callback, and handles emergency routing.

Parameter	Type	Description
`caller_phone`required	`string`	The caller's phone number.
`emergency_protocol`	`boolean`	Whether to enable emergency call routing.Default: `false`

end_of_day_report

Compiles an end-of-day summary including calls handled, appointments booked, leads captured, and any issues that need attention.

Custom Prompts

Create custom prompts in the FrontDeskOS dashboard to handle scenarios specific to your business. Custom prompts can reference any tools and resources available in your workspace.

Using Prompts Programmatically

Listing and using promptstypescript

// List all available prompts
const prompts = await mcp.listPrompts();

// Get a specific prompt with arguments
const prompt = await mcp.getPrompt("handle_incoming_call", {
  caller_phone: "+1-555-0123"
});

// The prompt contains structured messages for the AI
console.log(prompt.messages);
// [{ role: "user", content: { type: "text", text: "..." } }]