Resources · n8n integration

n8n Microsoft Agent 365 Trigger nodeConfigure Microsoft Agent 365 Trigger in n8n.

The n8n Microsoft Agent 365 Trigger node starts a workflow with its agent already inside. Four head parameters shape the run: a System Prompt that sets behavior, a switch for Microsoft Work IQ tools, an output format toggle and an Options collection. For teams putting an agent to work on Microsoft 365 data.

Try the Microsoft Agent 365 Trigger node on n8nSee the 4 parameters

Verified Trustpilot reviews · AI, automation & growth agency

Why automate

What does the n8n Microsoft Agent 365 Trigger node actually do?

It opens a workflow and runs an agent, meaning a piece of logic that reads a request, decides what to do and answers. Instead of wiring a trigger and an agent separately, this node carries both: the System Prompt tells the agent how to behave, and one switch gives it access to Microsoft tools like Calendar, Email and OneDrive. It ships inside the n8n AI package, so there is nothing to install.

First scenario, an internal assistant that answers scheduling questions. You write a System Prompt that describes the tone and the scope, you turn on Enable Microsoft Work IQ Tools for A365, and you narrow Tools to Include down to Calendar and Mail. The agent reads what it needs and replies, without a separate node for each lookup.

Second scenario, a document helper. Keep the same node, pick OneDrive, SharePoint and Word under Tools to Include, and let the workflow continue into whatever comes next, a message posted to Slack or a row appended in Google Sheets.

Third scenario, a structured hand off. Turn on Require Specific Output Format when the next node expects fields rather than a paragraph, then route the result into Gmail or into a system of your own. The agent stops improvising the shape of its answer.

When to pick something else: if the job is one call to an API and no reasoning is involved, the HTTP Request node is the shorter path. An agent earns its place when the steps are not known in advance, and a single call does not need one.

Known limits. The catalog lists this node at version 1, with 4 head parameters and no sub-node port, so a chat model, a memory or a tool is not plugged onto it the way it is on other AI nodes. The Microsoft tools need the matching permissions on the Microsoft account, and the model provider bills its own API calls under its own terms. If you are weighing platforms first, read the n8n review, and the n8n training covers the basics of expressions and credentials.

Connect

What do you connect before the first run?

  1. 01

    Open the Credentials menu

    In n8n, a credential is a saved set of access details that nodes reuse. Create it once from the Credentials menu, and every workflow on the instance can point at it, on n8n Cloud as well as on a self-hosted instance. You do not repeat the setup for each workflow, and rotating a key later happens in one place.

  2. 02

    Grant the Microsoft permissions

    The Microsoft tools are gated on the account, not on n8n. Enable Microsoft Work IQ Tools for A365 only works when the Microsoft account carries the right permissions for the tools you select, Calendar, Mail, OneDrive or the others. Sort that out with whoever administers the tenant before you start debugging the node itself.

  3. 03

    Add the node and run once

    Drop the node in an empty workflow, write a short System Prompt, leave the Options collection alone for now and execute. One run tells you whether the credential resolves and whether the agent answers. Keep the first prompt narrow, since a broad one makes it harder to read what went wrong.

Parameters

What does each parameter change?

The Microsoft Agent 365 Trigger node has 4 parameters. For each one: the node as you configure it in n8n, what the parameter changes, and our field notes.

01

System Prompt

systemPrompt

What you see in n8n

Notes & use cases

This is where the agent gets its instructions: who it is, what it may answer, how short to keep it. Nothing is required here, but an empty prompt leaves the agent guessing at scope.

Key parameters

  • System Prompt: free text, with a placeholder that shows the intended shape, e.g. You are a friendly assistant that helps people find a weather forecast for a given time and place.
Use cases
a support assistant told to answer in two sentences and to say when it does not know, rather than one that writes an essay per request.
02

Enable Microsoft Work IQ Tools for A365

useMcpTools

What you see in n8n

Notes & use cases

Flip this on and the agent may call Microsoft MCP tools such as Calendar, Email and OneDrive while it works on a task. It needs the matching permissions on the Microsoft account, so the switch alone is not enough.

Key parameters

  • Enable Microsoft Work IQ Tools for A365: a boolean, off by default in the sense that the tool fields below only appear once it is on.
  • Tools to Include: all (All) hands the agent everything, selected (Selected) opens the list.
  • Tools to Include: the named list, mcp_CalendarTools (Calendar), mcp_MailTools (Mail), mcp_TeamsServer (Teams), mcp_SharePointRemoteServer (SharePoint) among others.
Use cases
an agent that may read the calendar but must never touch mail, done with Selected rather than All.
03

Require Specific Output Format

hasOutputParser

What you see in n8n

Notes & use cases

By default the agent answers in prose. Turn this on when the node after it needs predictable fields, and n8n asks for a defined output shape instead of free text. An output parser, in n8n vocabulary, is what turns a model answer into structured data.

Key parameters

  • Require Specific Output Format: a boolean, nothing else to fill in at this level.
Use cases
a ticket classifier whose answer feeds a branch, where a sentence would break the comparison and a field will not.
04

Options

options

What you see in n8n

Notes & use cases

A collection you open with Add Option, holding the two settings that are rarely touched on the first run and matter a lot on the tenth.

Key parameters

  • Max Iterations: the maximum number of iterations the agent will run before stopping, your safety net against a loop that never resolves.
  • Welcome Message: the opening line shown to the person on the other side, e.g. Hello! I'm here to help you!
Use cases
capping iterations on an agent with many tools enabled, so a confused run ends instead of grinding on.
Need help

Need help automating Microsoft Agent 365 Trigger with n8n?

A person reads every message.

FAQ

Questions people ask next

01Is the n8n Microsoft Agent 365 Trigger node free on Cloud and self-hosted?
Yes on the n8n side. The node belongs to the n8n AI package, which ships with n8n, so there is nothing to install and no extra charge, whether you run n8n Cloud, the hosted offer, or a self-hosted instance under the Community Edition and its Sustainable Use license. A workflow behaves the same way in both cases. What does carry a cost sits outside n8n: the model provider bills its own API calls according to its own terms, and n8n adds nothing on top of that. Microsoft access follows your own agreement with Microsoft.
02What do you need for it to work?
A credential, created once from the Credentials menu in n8n and reused across every workflow on the instance. Beyond that, the Microsoft side matters: the tools behind Enable Microsoft Work IQ Tools for A365 require the appropriate permissions on your Microsoft account, and the node cannot grant those for you. Then come the parameters themselves. None of the 4 is mandatory, so the node runs with defaults, but a System Prompt is what makes the run useful rather than generic. Detailed credential documentation was not available when this page was written.
03What are the limits of this node?
The catalog lists version 1 with 4 head parameters, and that is the whole surface: a System Prompt, the Microsoft tools switch, an output format toggle and an Options collection with Max Iterations and Welcome Message. There is no sub-node port on it, so you do not attach a chat model, a memory or a tool the way you would elsewhere in the AI package. Tool availability depends on Microsoft permissions rather than on n8n. Older workflows may show an earlier node version with fewer options than what is described here.
04What should you set first, and what can wait?
Start with the System Prompt. It costs one minute and it decides most of the behavior, since scope and tone come from there rather than from the other settings. Next, decide on tools: leave the Microsoft switch off while you test the prompt, then turn it on and use Selected rather than All so the agent only reaches what it should. Require Specific Output Format comes into play once you know which node consumes the answer. Options can wait, with one exception, Max Iterations, worth setting as soon as several tools are enabled.
05n8n or Make for this kind of agent?
It depends on four things, and price is only one of them. Hosting: n8n runs self-hosted through Docker or npm, or on n8n Cloud, while Make is hosted only. Data control follows from that, since a self-hosted instance keeps the run and its data on infrastructure you own. Cost model differs too, as Make bills per operation. Visual logic is the last criterion, and that one is a matter of habit more than quality. Neither platform is wrong here. The answer depends on which of those four weighs most for your team.
Hack'celeration Lab

Get our weekly integration tips.

No spam. Unsubscribe anytime.

Same platform

Other docs you might need on this stack.

Browse all integrations