Example prompts

Once the Dime.Scheduler MCP serverMCP serverThe Model Context Protocol endpoint that exposes scheduling operations as tools to AI assistants. It is authenticated with API keys and translates tool calls into Dime.Scheduler API calls. is connected to your assistant, here are prompts you can try. They are written the way an end user would phrase them - a dispatcher, a planner, or a field service manager - not a developer.

Each example notes the tools the assistant is likely to call behind the scenes, so you can spot mistakes when a result feels off.

Dispatcher's daily check

"What are the open tasks today, and who is available to do them this afternoon?"

The assistant typically:

1. Calls get_open_tasks for the unplanned work.
2. Calls get_recommendations for each task, passing the taskTaskA unit of work that belongs to a job. It appears in the open task list until it is scheduled to a resource.'s skillsSkillA qualification a resource holds and a task can require, typically modeled as a filter group. Skills drive resource filtering, the project team view and the solver’s choice of who can do what. and a duration.
3. Summarizes the matches in a list.

Scheduling a known job for a known resource

"Schedule task TASK-001 for John Smith tomorrow at 2 PM for 2 hours."

The assistant has everything it needs to call create_appointment directly:

• taskNo: TASK-001
• resourceDisplayNames: John Smith
• startDateTime / endDateTime derived from "tomorrow at 2 PM" + 2 hours
• timeZone from the user's profile

It may first call parse_relative_time to resolve "tomorrow".

Finding a free slot

"When is John Smith available for a one-hour meeting this week?"

A simple find_available_slots call:

• resourceDisplayName: John Smith
• durationInMinutes: 60
• startDate / endDate derived from "this week"

Discovering the right resource

"Who is available next Wednesday morning for a 90-minute electrical job in Antwerp?"

This is a discovery question, so the assistant picks get_recommendations with:

• requirements: electrical work
• duration: PT1H30M
• locationAddress: Antwerp
• locationCountry: BE
• startDate / endDate for next Wednesday morning

The result is a ranked list of candidates, each with proposed slots.

Moving an appointment

"Move tomorrow's installation for Jane Smith to Friday at 9 AM."

The assistant calls reschedule_appointment with:

• subject: installation (or taskNo if it can find one)
• resourceDisplayName: Jane Smith
• currentDay: tomorrow
• newDay: Friday, newTime: 09:00

If the assistant accidentally reaches for get_recommendations here, that's a sign the prompt or the system message needs to make the discovery-vs-reassignment distinction clearer.

Inspecting the planning

"What's on Jane Smith's calendar next week?"

A direct call to get_resource_planning with the resolved date range.

Bulk look-up

"Show me all 'site visit' appointments next month, grouped by resource."

The assistant calls search_appointments with searchTerm: site visit and groups the results client-side. Increase limit if the time range is wide.

Validating a single time

"Is Jane Smith free on March 12th from 10 to 11 AM?"

A single check_time_slot_availability call. The response will include any conflicting periods if she isn't available.

Cancellation

"Cancel the 'Customer onsite' appointment for John Smith on Friday."

delete_appointment with subject, resourceDisplayName, and a resolved startDateTime for Friday.

Tips for better results

• Be specific about resourcesResourceAn entity that can carry out work - a person, vehicle, tool, or room - that you schedule on the planning board.. Display names are the unique identifier - "John" might be ambiguous, "John Smith" usually isn't.
• Be explicit about time zones if you work across regions, especially when scheduling for a customer in a different zone than the resource.
• Mention skills, not job titles. "Plumbing" and "electrical work" resolve cleanly; "the plumbing guy" may not.
• Ask the assistant to confirm before destructive actions. A short system instruction like "Always confirm before creating, updating, rescheduling, or deleting" prevents accidental changes.