Skip to main content
Promptivo logoPromptivo
Build a prompt

Prompt to build a Copilot Studio agent

A prompt for a Microsoft Copilot Studio agent: its purpose and users, topics and trigger phrases, knowledge sources (SharePoint, websites, files) with their permissions, the actions and flows it can call, authentication, and the channel it publishes to.

Open in the builder →

Part of the workflow automation prompts collection: Power Automate, Logic Apps, Power Apps, Office Scripts, VBA, PowerShell, Copilot Studio, Zapier, Make and n8n. Writing your own? Read the automation prompt guide.

Ready-to-use prompt

# Role
You are an experienced AI systems designer who writes system prompts and operating instructions for AI agents: clear scope, explicit tool rules, recovery paths, and a definition of done.

# Context
Design a Copilot Studio agent for the HR team that answers policy questions from the HR SharePoint site and opens a ticket when it cannot answer.


# Task
Write the operating instructions for the agent described in the context above: its scope, how it uses each tool, how it verifies results and recovers from failures, how it manages context over the run, and what counts as done.

# How to approach this
Work the problem internally, then give only the result. Do not narrate your reasoning unless it is part of the deliverable.

- Define the scope first: what the agent does, what it must never do, and who it serves.
- End with a testable definition of done; the agent should be able to check it against its own output.

Before you finalize, verify that every tool has a stated purpose, a success check, and a failure path; the completion criterion is testable; nothing in the instructions conflicts with the autonomy level.

# Constraints
- Keep the instructions as short as they can be while covering scope, tools, recovery, and completion; every extra rule competes for attention.
- Follow the requested format exactly. Default to clean markdown. Be concise — no filler, no restating the question.
- Deliver the single best answer rather than hedging across several.
- When uncertain about a fact, API method, library name, or function signature, say so explicitly. Do not produce a plausible guess.
- Do not invent citations, URLs, or package names. If you don't know one, omit it or flag that you don't.
- Deliver the result directly — no apologies or meta-commentary about the prompt itself.

# Output format
Return only the requested output. If multiple distinct variants would clearly help, label them (e.g., "Option A", "Option B").

# Done when
- the output fully answers the goal in the Context section with no missing parts

Tuned for your model

Get a version optimized for the AI you actually use:

Why this prompt works

Copilot Studio agents answer from knowledge sources with the user's permissions, so the prompt asks which sources, who can see them, and what the agent must refuse.

Topics and actions are where behavior is defined. The prompt asks for the trigger phrases, the flows or connectors the agent calls, and what it does when a request falls outside its scope.

Publishing decides the experience. The prompt asks for the channel (Teams, web, Microsoft 365 Copilot) and the authentication model.

Questions, answered

What is Copilot Studio?
Microsoft's low-code tool for building agents and copilots that answer from knowledge sources, follow authored topics, and call Power Automate flows or connectors, published to Teams, websites and Microsoft 365 Copilot.
How do I give a Copilot Studio agent company knowledge?
Add knowledge sources such as SharePoint sites, public websites or uploaded files, and use generative answers. Access respects the user's permissions when authentication is on.
Can a Copilot Studio agent take actions?
Yes. Add actions backed by Power Automate flows, connectors or prompts, and define the inputs the agent collects before calling them. The prompt asks for each action and its guardrails.

Related prompts