Microsoft Teams - Prisme.ai

The Microsoft Teams app provides read/write access to Microsoft Teams through the Microsoft Graph REST API (v1.0): teams and channels, channel messages and their reply threads, one-to-one and group chats, and online meetings with their transcripts, recordings and attendance reports. It can be consumed three ways: as a remote MCP server that Agent Factory agents call as tools, as a Builder app whose instructions you call directly from DSUL, and as a chat connector — an inbound Teams bot that hands each incoming message to your own automation. The MCP surface groups every operation into four entity tools (teams, channelMessages, chats, meetings), each driven by an action argument, and runs in the tenant app-instance context (it resolves the installing workspace’s own credentials). Authentication is per-user and supports several modes:

Per-user OAuth2 (central client) (oauthCentral, recommended): one Microsoft Entra OAuth Application is registered once by the platform maintainer; every end user signs in with their own Microsoft account. Nothing to register per tenant: each workspace just installs the app and clicks Connect.
Per-user OAuth2 (tenant client) (oauth): paste your own Entra application (client) ID/secret in the connector config app. Each user signs in with their own account against your client (PKCE authorization-code flow).
Application (client credentials) (clientCredentials): an Entra app with application Graph permissions and admin consent, acting without an interactive sign-in. Best for back-office / service automations and content indexing.
Direct access token (accessToken): a caller-managed Microsoft Graph access token, used as-is with no exchange.

Conversations

Browse joined teams and channels, read channel messages and their reply threads, read and post one-to-one and group chats, and inspect team, channel and chat rosters

Meetings & transcripts

Find meetings through the calendar, resolve a join URL into a meeting, schedule and update meetings, and read transcripts, recordings metadata and attendance reports

Chat connector

Register a Teams bot whose messages are signature-verified and handed to your automation, so people can talk to an agent from a Teams conversation

Who is this for?

This connector is used by three different roles. Jump to the section that matches yours; each one is self-contained.

Agent builder

You build agents in Agent Factory and want them to read Teams conversations and meeting transcripts. → Agent builder tab.

Platform admin

You run the platform and set up the shared Microsoft Entra OAuth client once for everyone. → Platform admin setup accordion below.

Workspace builder

You write Builder automations (DSUL) that call Teams operations, or run the inbound bot. → Workspace builder and Chat connector tabs.

Prerequisites (Microsoft side)

A Microsoft 365 tenant with Teams, and an Azure / Microsoft Entra ID tenant where you can register an application (or a maintainer who already did).
An app registration in Microsoft Entra ID (Azure portal → App registrations), with a client secret created under Certificates & secrets.
Microsoft Graph delegated permissions granted on the app (for the per-user OAuth modes), covering the resources you intend to use:
- Teams & channels: Team.ReadBasic.All, Channel.ReadBasic.All, TeamMember.Read.All
- Channel messages: ChannelMessage.Read.All (add ChannelMessage.Send to post)
- Chats: Chat.Read, ChatMessage.Read (or Chat.ReadWrite + ChatMessage.Send to post)
- Meetings: Calendars.Read (meeting discovery goes through the calendar), OnlineMeetings.Read (or OnlineMeetings.ReadWrite), OnlineMeetingTranscript.Read.All, OnlineMeetingArtifact.Read.All
- Sign-in / identity: openid, profile, offline_access, User.Read
For the application (client-credentials) mode instead, grant the equivalent application permissions and have a Global Administrator grant admin consent. Reading channel messages with application permissions additionally requires Microsoft’s protected API approval.
For the chat connector only: an Azure Bot resource with the Microsoft Teams channel enabled (see the Chat connector tab).

The OAuth scopes requested by the read-only preset are:

offline_access
https://graph.microsoft.com/User.Read
https://graph.microsoft.com/Team.ReadBasic.All
https://graph.microsoft.com/Channel.ReadBasic.All
https://graph.microsoft.com/TeamMember.Read.All
https://graph.microsoft.com/ChannelMessage.Read.All
https://graph.microsoft.com/Chat.Read
https://graph.microsoft.com/ChatMessage.Read
https://graph.microsoft.com/Calendars.Read
https://graph.microsoft.com/OnlineMeetings.Read
https://graph.microsoft.com/OnlineMeetingTranscript.Read.All
https://graph.microsoft.com/OnlineMeetingArtifact.Read.All

Platform admin (Governance): one-time platform setup

Goal: two one-time tasks: (1) configure the shared central Microsoft Entra OAuth client so every workspace lets its users sign in with their own Microsoft account, and (2) expose Microsoft Teams as a reusable capability in AI Governance so agent builders can pick it without pasting endpoint URLs.

1. Configure the connector

In Microsoft Entra ID → App registrations > New registration, register a Web application. Add a Redirect URI pointing at the core workspace callback:

<api-url>/workspaces/<workspace-id>/webhooks/oauthCallback

Microsoft Entra rejects redirect URIs that contain a colon in the path, so use the raw workspace id of the core Microsoft Teams workspace, not the slug:microsoft-teams-next form used by other connectors. Copy the exact URI from the connector’s configuration app, which renders it for you.

Add the Microsoft Graph delegated permissions listed in Prerequisites, create a client secret under Certificates & secrets, and note the Application (client) ID, the secret value, and the Directory (tenant) ID (or use common / organizations for multi-tenant).

Enter the credentials through the configuration app

Open the central microsoft-teams-next workspace and launch its Configuration app: <studio>/apps/microsoft-teams-next, also linked as Configuration app on the installed instance. Switch to the maintainer view and follow the in-app instructions to paste the Client ID, Client Secret, Tenant and (optionally) scopes; the app stores them in the core workspace’s secrets for you. Do not edit Studio’s raw Secrets by hand. These credentials stay in the microsoft-teams-next workspace and are never exposed to tenants or end users; token exchange is proxied through the core centralTokenExchange webhook so the client secret never leaves the core workspace.

Tell workspaces to use the central client

Each consuming workspace selects auth mode oauthCentral in the connector configuration app (no client id/secret to enter on their side). Their users then just click Connect.

2. Declare the capability in AI Governance

Generic connectors (broad tool surfaces meant to be shared across many agents, like Microsoft Teams) are best exposed as a named capability in AI Governance. Agent builders then enable that capability on their agents instead of pasting a raw MCP endpoint.

Open AI Governance > Capabilities

Create (or edit) the Microsoft Teams capability.

Point it at the MCP endpoint

Set the capability’s MCP server URL to the connector’s MCP Endpoint, and set its Scope to:

context_id,agent_id,user_id

The agent_id in the scope is what lets the connector identify and authorize the calling agent.

Make it available to agent builders

Once created, the capability appears in the capability picker for agent builders in your organization, who enable it on their agents. Access to the catalog follows your organization’s existing roles; there is no per-capability role grant.

Smoke-test

From an agent that has the capability, in a workspace configured for oauthCentral, trigger any tool. The user is prompted to connect once (Microsoft sign-in); subsequent calls reuse the stored token transparently and refresh it automatically.

The connector’s configuration app also offers a one-click Add to catalog button (workspace owner / admin only) that publishes the Microsoft Teams capability to the organization-wide Capabilities catalog for you, the easiest way to expose it to agent builders without hand-editing Governance.

Declaring the capability makes the connector available; it does not by itself authorize a specific agent. This connector follows the tenant-context model: which agents may actually call it is gated per-workspace by the authorized-agents allowlist in the configuration app (see the Workspace builder tab). There is also no OAuth auth-config JSON to attach in Governance: connect / status / disconnect are handled by the connector’s own webhooks, wired automatically.

Agent builder (Agent Factory)
Workspace builder (DSUL)
Chat connector (Teams bot)

Agent builder

Goal: let an agent you build in Agent Factory read Teams conversations, post messages, and pull meeting transcripts through MCP tools.

Before an agent can call the connector, a Workspace builder must have installed and configured the Microsoft Teams app in a workspace (see the Workspace builder tab), and, for the central OAuth mode, a Platform admin must have provisioned the shared Microsoft Entra OAuth client (see the Platform admin setup accordion above).

This connector runs in the tenant app-instance context: your agent is authorized two ways at once: it is identified by the agent_id that Agent Factory injects through the capability Scope, and that agent must appear in the connector’s authorized-agents allowlist (managed in the configuration app). The Microsoft Graph access token itself is resolved server-side from the configured auth mode.There are two ways to wire it up. Pick based on how much isolation you need.

Option A: Enable the shared capability from the catalog

The fastest path: a Platform admin has already published a Microsoft Teams capability (see the Platform admin setup accordion above, §2), so you just pick it from the catalog. A workspace owner can publish it in one click with the config app’s Add to catalog button.

Open your agent in Agent Factory

Open the agent you want to extend and go to its capabilities / tools.

Add the Microsoft Teams capability

Browse the capability catalog, select Microsoft Teams, and enable it. The MCP endpoint URL and the Scope (context_id,agent_id,user_id) are already wired by the admin: nothing to paste, and the shared instance accepts every agent, so there is no allowlist step on your side.

Connect a Microsoft account (OAuth modes)

On the first tool call, an unconnected user is prompted to sign in; Agent Factory surfaces a connect_url. Application (clientCredentials) and accessToken modes need no per-user sign-in.

Convenient, but your agent runs against a shared, platform-managed instance: its Microsoft credentials are owned by someone else and the instance accepts every agent that is granted the capability. Prefer Option B for anything beyond quick experiments.

Option B: Run it from your own workspace (recommended)

For production agents, install the connector in your own workspace and point the agent at that workspace’s MCP endpoint.

Prefer this mode for security. Because the MCP runs in your app-instance context, the Microsoft credentials, the per-user OAuth tokens and the authorized-agents allowlist are all scoped to your workspace, not shared platform-wide. You decide exactly which agents may call it and which Microsoft account / auth mode backs them, and a misconfiguration elsewhere can never expose your Teams conversations. The shared catalog capability (Option A) is a broad surface many agents can reach; your own workspace is an isolated, least-privilege boundary.

Install and configure the connector in your workspace

Follow the Workspace builder tab: install Microsoft Teams in your workspace, open its Configuration app, choose the auth mode and connect a Microsoft account.

Allowlist your agent

In that workspace’s config app, open Authorized agents and tick your agent (the Install capability button does this for you).

Add the MCP capability to your agent

In your agent, add a capability pointing at your workspace’s MCP Endpoint URL, and set its Scope to:

context_id,agent_id,user_id

The agent_id is what lets the connector identify and authorize your agent; without it, every call is rejected with an explicit “agent could not be identified” message. This Scope is separate from the Microsoft Graph OAuth scopes.

Connect a Microsoft account (OAuth modes)

On the first tool call, the user is prompted to sign in (or uses Connect in the config app).

Brief the agent in its system prompt

Whichever option you pick, wiring the capability is not enough: the agent must know the MCP exists and when to use it. Copy-pasteable starter:

You have access to the Microsoft Teams MCP server (tools: teams, channelMessages, chats, meetings). Each tool takes an `action` argument. Use it whenever the user asks about Teams teams or channels, channel discussions, chats, or meetings and their transcripts. Resolve ids before reading content: use `teams` (action list, then listChannels) to get teamId and channelId, then `channelMessages`. A channel conversation is two levels deep: `list` returns the thread starters and `listReplies` returns the body of a discussion, so read both before summarizing a thread. To reach a meeting transcript, call `meetings` action list (this reads the calendar, since Graph has no "list my meetings"), then `resolve` on the joinUrl to get a meetingId, then listTranscripts and getTranscript. Prefer calling a tool over guessing, and confirm with the user before posting anything (send, reply) or creating or updating a meeting.

Legacy AI Knowledge agents (no native MCP picker): add the connector under Advanced > Tools > MCP and paste the MCP Endpoint URL. The agent still has to be allowlisted in the config app and its identity propagated so the connector can read its agent_id.

Restricting to read-only (least privilege). The Teams tools cover both reads and writes (posting a channel message, replying, sending a chat message, creating or updating a meeting). The connector requests exactly the scopes configured in the Scopes field of the configuration app, which offers a read-only preset; writes are only possible if you widen it. To keep the connector read-only, use the read preset listed in Prerequisites — in particular leave out ChannelMessage.Send, Chat.ReadWrite, ChatMessage.Send and OnlineMeetings.ReadWrite. Any write attempt is then rejected by Microsoft Graph with 403 Forbidden, while the connection itself stays healthy because every connection check is a read.With the central client (oauthCentral) you do not create your own Entra app: you keep oauthCentral and simply enter the read-only scopes, which override the central default — provided the platform’s central app exposes those read permissions. The scope list is set at the workspace level (a workspace editor can widen it again); for a hard guarantee, back the connection with an Entra app limited to read permissions.

Available Tools

Each tool takes an action argument selecting the concrete operation, plus the per-action parameters.

Tool	Description
`teams`	Teams the user belongs to, their channels and their rosters. Actions: `list`, `get`, `listChannels`, `getChannel`, `listMembers`, `listChannelMembers`.
`channelMessages`	Messages posted in a channel and their replies. A channel conversation is two levels deep: `list` returns thread starters, `listReplies` the body of a discussion. Actions: `list`, `get`, `listReplies`, `getReply`, `send`, `reply`.
`chats`	One-to-one and group conversations, outside any channel. A chat is flat: there are no replies. Actions: `list`, `get`, `listMessages`, `getMessage`, `send`, `listMembers`.
`meetings`	Online meetings: discovery, scheduling, transcripts, recordings, attendance. Actions: `list`, `resolve`, `get`, `create`, `update`, `listTranscripts`, `getTranscript`, `listRecordings`, `listAttendanceReports`, `getAttendanceReport`.

Two Graph behaviours worth knowing before you prompt an agent. A meeting is discovered through the calendar: Graph has no “list my online meetings” collection, so meetings list reads calendar events and resolve turns a joinUrl into the meetingId every other action needs. And a transcript is WebVTT text, not JSON: getTranscript returns speaker-tagged plain text, truncated on a line boundary (maxTranscriptCharacters, 20000 by default) so a long meeting cannot exhaust the agent’s context window.

Output Formats

Every tool accepts an outputFormat argument that controls the MCP response shape:

verbose (default): a human-readable text view, optimized for LLM consumption.
structured: concise machine-readable JSON in structuredContent.
both: the structured payload, with its JSON also rendered as text.

Knowledge Resources

The connector also exposes Teams conversations as MCP Resources, so a knowledge base can index them. The tree is flattened to the discussion: a team and a channel are containers, and a thread — a root message plus its replies — is the document. Each thread’s revision folds in the timestamp of the newest reply and the reply count, because the root message’s own lastModifiedDateTime does not move when someone answers; without that, an answered question would never be re-indexed. Access is checked per reader, at that reader’s own identity.Channel files are not indexed here: they live in the team’s SharePoint site and belong to the SharePoint connector.

Tool Details

teams

{
  "name": "teams",
  "arguments": {
    "action": "listChannels",
    "teamId": "19:abc...@thread.tacv2"
  }
}

Parameter	Required	Description
`action`	Yes	One of `list`, `get`, `listChannels`, `getChannel`, `listMembers`, `listChannelMembers`.
`teamId`	For every action except `list`	Team id.
`channelId`	For getChannel/listChannelMembers	Channel id.
`userId`	No (list only)	Target user; omit for the signed-in user.
`top` / `select` / `filter` / `skiptoken`	No	OData query options and pagination (`top` 1..100, default 20).

list accepts no OData query option: Microsoft Graph answers 400 Query option 'Top' is not allowed on joinedTeams rather than ignoring it. It returns every team the identity belongs to in one response, so there is nothing to page.

channelMessages

{
  "name": "channelMessages",
  "arguments": {
    "action": "listReplies",
    "teamId": "19:abc...@thread.tacv2",
    "channelId": "19:def...@thread.tacv2",
    "messageId": "1700000000000"
  }
}

Parameter	Required	Description
`action`	Yes	One of `list`, `get`, `listReplies`, `getReply`, `send`, `reply`.
`teamId`	Yes	Team id.
`channelId`	Yes	Channel id.
`messageId`	For get/listReplies/getReply/reply	Root message id.
`replyId`	For getReply	Reply id.
`top` / `select` / `expand` / `skiptoken`	No	OData query options; `top` is clamped to 1..50 by Graph on channel messages. `expand=replies` inlines a whole thread in one call.
`body`	For send/reply	A chatMessage resource, e.g. `{ body: { contentType: 'html', content: '...' } }`.

chats

{
  "name": "chats",
  "arguments": {
    "action": "listMessages",
    "chatId": "19:abc...@unq.gbl.spaces",
    "top": 20
  }
}

Parameter	Required	Description
`action`	Yes	One of `list`, `get`, `listMessages`, `getMessage`, `send`, `listMembers`.
`chatId`	For every action except `list`	Chat id.
`messageId`	For getMessage	Message id.
`userId`	No (list only)	Target user; omit for the signed-in user.
`top` / `select` / `expand` / `filter` / `orderby` / `skiptoken`	No	OData query options; `top` is clamped to 1..50 by Graph on chat messages.
`body`	For send	A chatMessage resource, e.g. `{ body: { contentType: 'text', content: '...' } }`.

Chats are private conversations: everything here is read at the connected user’s own identity, and a chat they have left is no longer readable.

meetings

{
  "name": "meetings",
  "arguments": {
    "action": "resolve",
    "joinUrl": "https://teams.microsoft.com/l/meetup-join/19%3ameeting_..."
  }
}

Parameter	Required	Description
`action`	Yes	One of `list`, `resolve`, `get`, `create`, `update`, `listTranscripts`, `getTranscript`, `listRecordings`, `listAttendanceReports`, `getAttendanceReport`.
`meetingId`	For every action except list/resolve/create	Online meeting id; obtain it with `resolve`.
`joinUrl`	For resolve	The meeting join URL, as returned by `list`.
`transcriptId`	For getTranscript	Transcript id, from `listTranscripts`.
`reportId`	For getAttendanceReport	Attendance report id.
`from` / `to`	No (list only)	ISO 8601 window bounds, e.g. `2026-01-01T00:00:00Z`.
`userId`	No	Target user; omit for the signed-in user. Application-mode calls must set it, since there is no `me` without a signed-in user.
`maxTranscriptCharacters`	No (getTranscript only)	Truncate the WebVTT beyond this many characters; default 20000, clamped 2000..120000.
`body`	For create/update	An onlineMeeting resource, e.g. `{ startDateTime, endDateTime, subject }`.

The usual chain is list → resolve(joinUrl) → listTranscripts → getTranscript.

Workspace builder

Goal: install the connector in a workspace, configure authentication and the agent allowlist, and call Teams operations from your automations.

Installation

Go to Apps in your workspace
Search for Microsoft Teams and install it
Open the Configuration app (the link auto-populated on install) to choose the auth mode, provide credentials, connect, and allow the agents that may call the connector

Configuration

Field	Description
Configuration app	Auto-populated on install; open this link to configure authentication (mode + credentials), connect a Microsoft account (OAuth modes), manage the authorized-agents allowlist, and register the Teams bot.
Microsoft Graph API version	`v1.0` (default) or `beta`.
Knowledge-sync	Which workspaces may run a delegated Resource access check on behalf of a named reader. Defaults to the platform’s knowledge-sync workspace.

The configuration app drives everything; the app instance itself has no per-field credential form. It is organised in three tabs — Authentication, Agents (MCP endpoint, catalog entry and the authorized-agents allowlist) and Teams bot — and from the first one you pick one of:

Auth mode	What you provide	Best for
`oauthCentral`	Nothing; just click Connect	Workspaces on a platform where the admin set up the central client
`oauth`	Your own Entra Application (client) ID + Secret (+ tenant; redirect URI `<api-url>/workspaces/<your-workspace-id>/webhooks/oauthCallback`)	You manage your own Microsoft Entra OAuth app
`clientCredentials`	Entra client ID + secret + tenant with admin-consented application Graph permissions	Server-to-server / back-office, content indexing, no interactive sign-in
`accessToken`	A pre-minted Microsoft Graph access token	Short-lived / caller-managed tokens

Microsoft Entra rejects a colon in the redirect URI path, so for the oauth mode use your workspace’s raw id in the redirect URI, not the slug: form. The configuration app shows the exact URI to register.

Credentials are provisioned into the workspace’s Secrets by the onInstall flow and resolved server-side. Always set them through the configuration app rather than Studio’s Secrets screen. The agent allowlist (Authorized agents) gates which Agent Factory agents may call the MCP endpoint; see the Agent builder tab.

Available Instructions

Every instruction resolves credentials from the workspace configuration. Operations are grouped by Teams resource; the MCP server exposes the same operations behind the four entity tools (see the Agent builder tab). Instructions that read or write on behalf of a specific person accept an optional userId; omit it for the signed-in user.

Teams & channels

Instruction	Description	Returns
`listJoinedTeams`	List the teams the identity belongs to. Accepts no OData option.	`{ value: [{ id, displayName, description }] }`
`getTeam`	Get a team by `teamId`.	A team resource `{ id, displayName, visibility, … }`.
`listTeamMembers`	List the members of a `teamId`.	`{ value: [{ id, displayName, email, roles }] }`
`listChannels`	List the channels of a `teamId`.	`{ value: [{ id, displayName, membershipType, webUrl }] }`
`getChannel`	Get a channel by `teamId` + `channelId`.	A channel resource.
`listChannelMembers`	List the members of a channel.	`{ value: [ conversationMember ] }`

Channel messages

Instruction	Description	Returns
`listChannelMessages`	Root messages of a channel, newest first (`teamId`, `channelId`). `expand=replies` inlines each thread.	`{ value: [ chatMessage ], "@odata.nextLink" }`
`getChannelMessage`	Get one root message by `messageId`.	A chatMessage `{ id, body, from, createdDateTime }`.
`listChannelMessageReplies`	Replies to a root message — where the body of a discussion lives.	`{ value: [ chatMessage ] }`
`getChannelMessageReply`	Get a single reply by `replyId`.	A chatMessage.
`sendChannelMessage`	Post a new root message; `body` = `{ body: { contentType, content }, subject? }`.	The created chatMessage.
`replyToChannelMessage`	Reply to a root message; `body` = `{ body: { contentType, content } }`.	The created chatMessage.

Chats

Instruction	Description	Returns
`listChats`	Chats the identity takes part in, most recently updated first.	`{ value: [{ id, topic, chatType, lastUpdatedDateTime }] }`
`getChat`	Get a chat by `chatId`.	A chat resource.
`listChatMessages`	Messages of a chat, newest first. A chat is flat: no replies.	`{ value: [ chatMessage ] }`
`getChatMessage`	Get one message by `chatId` + `messageId`.	A chatMessage.
`sendChatMessage`	Post a message to a chat; `body` = `{ body: { contentType, content } }`.	The created chatMessage.
`listChatMembers`	List the participants of a chat.	`{ value: [ conversationMember ] }`

Meetings

Instruction	Description	Returns
`listMeetingEvents`	Calendar events carrying a Teams meeting, within `from`/`to`. This is how you find a meeting.	`{ value: [{ id, subject, start, end, onlineMeeting: { joinUrl } }] }`
`resolveOnlineMeeting`	Turn a `joinUrl` into the online meeting every other operation needs.	`{ value: [ onlineMeeting ] }`
`getOnlineMeeting`	Get a meeting by `meetingId`.	An onlineMeeting resource.
`createOnlineMeeting`	Schedule a meeting; `body` = `{ startDateTime, endDateTime, subject }`.	The created onlineMeeting, with `joinWebUrl`.
`updateOnlineMeeting`	Change a meeting; `body` = the fields to update.	The updated onlineMeeting.
`listMeetingTranscripts`	Transcript ids and dates for a meeting — not the text.	`{ value: [{ id, createdDateTime }] }`
`getMeetingTranscriptContent`	The transcript itself, by `transcriptId`.	WebVTT plain text (speakers + timestamps), truncated on a line boundary — not JSON.
`listMeetingRecordings`	Recording metadata; the video is never returned inline.	`{ value: [{ id, createdDateTime, recordingContentUrl }] }`
`listMeetingAttendanceReports`	Attendance reports of a meeting.	`{ value: [{ id, totalParticipantCount, meetingStartDateTime }] }`
`getMeetingAttendanceReport`	One report with its records: who joined, when, for how long.	A report with `attendanceRecords`.

Bot bridge

Instruction	Description	Returns
`sendTextMessage`	Reply with plain text (Markdown supported) in the Teams conversation an inbound activity came from. `originalActivity` = the activity received on `teams.bot.messageReceived`.	The Bot Connector response `{ id }`.

Returns shows the shape of the operation output (the underlying Microsoft Graph resource). getMeetingTranscriptContent is the exception: it returns WebVTT text, because Graph serves transcripts as text rather than JSON.

DSUL Examples

List joined teams and their channels:

- MicrosoftTeamsNext.listJoinedTeams:
  output: teams
- MicrosoftTeamsNext.listChannels:
    teamId: '{{teams.value[0].id}}'
  output: channels

Read a channel discussion, thread starters then replies:

- MicrosoftTeamsNext.listChannelMessages:
    teamId: '{{teams.value[0].id}}'
    channelId: '{{channels.value[0].id}}'
    top: 20
  output: messages
- MicrosoftTeamsNext.listChannelMessageReplies:
    teamId: '{{teams.value[0].id}}'
    channelId: '{{channels.value[0].id}}'
    messageId: '{{messages.value[0].id}}'
  output: replies

Post a message in a channel:

- MicrosoftTeamsNext.sendChannelMessage:
    teamId: '{{teams.value[0].id}}'
    channelId: '{{channels.value[0].id}}'
    body:
      body:
        contentType: html
        content: '<p>Deployment finished.</p>'
  output: posted

Find a meeting and read its transcript:

- MicrosoftTeamsNext.listMeetingEvents:
    from: '2026-01-01T00:00:00Z'
    to: '2026-01-31T23:59:59Z'
    top: 20
  output: events
- MicrosoftTeamsNext.resolveOnlineMeeting:
    joinUrl: '{{events.value[0].onlineMeeting.joinUrl}}'
  output: meeting
- MicrosoftTeamsNext.listMeetingTranscripts:
    meetingId: '{{meeting.value[0].id}}'
  output: transcripts
- MicrosoftTeamsNext.getMeetingTranscriptContent:
    meetingId: '{{meeting.value[0].id}}'
    transcriptId: '{{transcripts.value[0].id}}'
  output: transcript

Schedule a meeting:

- MicrosoftTeamsNext.createOnlineMeeting:
    body:
      subject: 'Quarterly review'
      startDateTime: '2026-02-10T09:00:00Z'
      endDateTime: '2026-02-10T10:00:00Z'
  output: created

Chat connector

Goal: let people talk to one of your agents from inside Teams — a one-to-one chat with a bot, or a bot mentioned in a channel.This is the connector’s third surface, and it works the other way round from the other two. There, you call Microsoft. Here, Microsoft calls you: an Azure Bot forwards every message to the connector, which verifies it and hands it to an automation you write. Nothing is routed to an agent automatically — you decide what answers, which is what lets you put an approval step, a filter or a router in front of it.

The bot is optional and independent from the rest: a workspace that only uses the tools and Knowledge Resources never needs one. Conversely, a bot needs no Microsoft Graph OAuth connection — it authenticates as the bot, not as a user.

1. Register an Azure Bot

Create the Azure Bot resource

In the Azure portal, create an Azure Bot resource. Choose a Multi Tenant or Single Tenant app type and let Azure create the Microsoft App ID, or reuse an existing app registration. Note the Microsoft App ID, and create a client secret for it.

Enable the Microsoft Teams channel

In the bot resource, open Channels and add Microsoft Teams. Without it, the bot exists but Teams will never deliver a message to it.

Point the messaging endpoint at the connector

In Configuration, set the Messaging endpoint to the URL shown in the connector’s configuration app:

<api-url>/workspaces/<your-workspace-id>/webhooks/<app-instance>.messagesEndpoint

The configuration app renders the exact value with a copy button; <app-instance> is the slug of the installed app instance.

Enter the bot credentials in the configuration app

Open the connector’s Configuration app and go to its Teams bot tab, then paste the Bot app ID and Bot app password (the client secret). Leave Bot tenant empty for a multi-tenant bot; set your Entra tenant id for a single-tenant one. Save.

Enter these in the configuration app, never in Studio’s Secrets screen. That screen serializes object values to strings, which would break not only the bot but the workspace’s Microsoft Graph credentials with it.

Install the bot in Teams

Package the bot as a Teams app (Developer Portal or an app manifest referencing the Microsoft App ID) and install it in the team or for the users who should be able to talk to it.

2. Answer incoming messages

Every verified activity is emitted as an event in your workspace. Write an automation that listens to it and replies.

Event	Payload	When
`teams.bot.messageReceived`	The full Bot Framework activity — `type`, `text`, `from`, `conversation`, `serviceUrl`, `id`, …	Every inbound activity that passed verification. This is the one to listen to.
`teams.bot.activityReceived`	A summary: `activityType`, `conversationId`, `fromId`, `fromName`	Same trigger, lighter payload — convenient for logging or metrics.
`teams.bot.rejected`	`reason`, `activityType`	An activity failed verification and was answered `401`. Useful for alerting; the caller is never told which check failed.

A minimal echo, to check the round trip end to end:

slug: replyInTeams
when:
  events:
    - teams.bot.messageReceived
do:
  - conditions:
      '{{payload.type}} != "message"':
        - break: {}
  - MicrosoftTeamsNext.sendTextMessage:
      text: 'You said: {{payload.text}}'
      originalActivity: '{{payload}}'

Passing the whole activity back as originalActivity is what tells the connector where to answer: the conversation, the message being replied to and the serviceUrl are all read from it.

3. Hand the message to an agent

To make the bot conversational, call your agent between the two steps and send its answer back:

slug: askAgentFromTeams
when:
  events:
    - teams.bot.messageReceived
do:
  - conditions:
      '{{payload.type}} != "message"':
        - break: {}
  - comment: Strip the @mention Teams prepends in a channel, so the agent sees the question only.
  - set:
      name: question
      value: '{{payload.text}}'
  - AI Knowledge.query:
      projectId: '{{config.agentId}}'
      question: '{{question}}'
    output: answer
  - MicrosoftTeamsNext.sendTextMessage:
      text: '{{answer.output.text}}'
      originalActivity: '{{payload}}'

Replace the middle step with whatever should answer — an agent, a search, a routing table, or a human approval before replying. Teams renders Markdown in the reply, so lists and links come through.

Bot Framework expects the webhook itself to answer immediately and empty; the reply travels back on a separate call. That is why the connector acknowledges Microsoft right away and your automation replies afterwards, and why a slow agent does not cause Teams to retry the message.

What the connector guarantees

The messaging endpoint is necessarily anonymous — Microsoft calls it from outside and carries no Prisme.ai identity — so the signed token it sends is the only gate, and it is checked before anything is emitted. An event here makes a bot speak to a real person in a real conversation, so an unverified payload would be a way for anyone to put words in the bot’s mouth. On each inbound activity the connector verifies:

the RS256 signature, against Microsoft’s published Bot Connector keys (cached, and re-fetched once if a key rotated);
the issuer and the audience (your bot’s app id);
the token’s validity window;
the serviceUrl carried by the activity, which matters because the reply posts the bot token to that URL: without this check a replayed activity could aim it at a host of the attacker’s choosing.

A failed check returns 401 without saying which one failed — telling an unauthenticated poster that would be free reconnaissance — and emits teams.bot.rejected so you can still see it. If Microsoft’s key endpoint cannot be reached, the connector answers 503 rather than trusting the activity, and Microsoft retries.

Message text is registered as sensitive before it can reach an event, so conversation content is redacted from the activity feed rather than logged in clear.

Limits

Plain text and Markdown only. Adaptive cards, typing indicators and attachments use the raw Bot Framework activity shape, which is not exposed as an instruction; sendTextMessage covers the conversational case.
No built-in agent routing. The connector delivers the message and sends the reply; choosing what answers is the automation you write, as in the example above.
The reply targets the originating conversation. Starting a conversation with a user who has never talked to the bot (proactive messaging) is not covered.

Error Handling

HTTP code	Meaning
`400`	Bad request: invalid parameters or malformed body. Note that some Teams endpoints reject OData options rather than ignoring them (`joinedTeams` refuses `$top`).
`401`	Unauthorized: the Microsoft Graph access token is missing, expired or revoked. The user must reconnect (OAuth modes). On the bot endpoint: the activity’s signature, issuer, audience, validity window or `serviceUrl` did not check out.
`403`	Forbidden: insufficient Graph scope/permission, missing admin consent (application mode), or the account is not a member of the team or chat.
`404`	Not found: the team, channel, chat, message or meeting id does not exist or is not visible to the account.
`429`	Throttled: Microsoft Graph rate limit; back off and retry (honor `Retry-After`).
`500` / `503`	Service error: transient; retry shortly. On the bot endpoint, `503` means Microsoft’s signing keys could not be fetched — the activity was deliberately not trusted.

Common Issues

“This agent is not authorized to use this connector”: The calling agent is not in the allowlist. Open the configuration app → Authorized agents → tick this agent (or enable Allow all agents) and Save. “The calling agent could not be identified”: The MCP capability Scope does not declare agent_id, so Agent Factory never injects the agent identity. Set the Scope to context_id,agent_id,user_id on the capability, then allow the agent in the config app. “Microsoft Teams is not connected for this user”: No per-user OAuth token. Open the configuration app (OAuth mode) and click Connect, or use the agent’s connect flow. “token refresh failed … must reconnect”: The stored refresh token was revoked or expired (Microsoft Entra invalidated it, or it aged past its limit). The connection is dropped automatically; the user must reconnect from the config app. “OAuth is not configured”: Neither a tenant OAuth client nor the central platform client is available. Set the client ID/secret/tenant in the config app, or ask the platform maintainer to provision the central Entra OAuth client. “Bot Framework credentials are not configured”: The bot app id or password is missing. Fill them in the configuration app’s Teams bot tab; do not set them from Studio’s Secrets screen. The bot receives nothing: Check three things in order — the Microsoft Teams channel is enabled on the Azure Bot resource, the messaging endpoint matches the URL shown in the config app exactly, and the bot app is actually installed in the team or chat. A teams.bot.rejected event means the message arrived but failed verification, which is a credential mismatch rather than a wiring problem. “Query option ‘Top’ is not allowed”: Some Teams collections reject OData options outright instead of ignoring them. listJoinedTeams returns every team in one response; do not pass top or a paging cursor to it. A meeting cannot be found: There is no “list my online meetings” endpoint in Graph. Discover meetings with listMeetingEvents (the calendar), then turn the event’s joinUrl into a meeting with resolveOnlineMeeting. Application mode cannot read channel messages: Reading Teams messages with application permissions requires Microsoft’s protected API approval on top of admin consent. Per-user OAuth is not subject to it. Microsoft Entra redirect URI rejected: Entra rejects a colon in the redirect URI path. Register the redirect URI with the workspace’s raw id (/workspaces/<workspace-id>/webhooks/oauthCallback), not the slug: form; copy the exact value from the config app.

External Resources

Microsoft Graph API

Official reference for the Teams, channels, chats, messages and online meetings REST APIs.

Tool Agents

Learn how Agent Factory agents consume MCP tools in Prisme.ai.

Conversations

Meetings & transcripts

Chat connector

​Who is this for?

Agent builder

Platform admin

Workspace builder

​Prerequisites (Microsoft side)

​1. Configure the connector

​2. Declare the capability in AI Governance

​Agent builder

​Option A: Enable the shared capability from the catalog

​Option B: Run it from your own workspace (recommended)

​Brief the agent in its system prompt

​Available Tools

​Output Formats

​Knowledge Resources

​Tool Details

​teams

​channelMessages

​chats

​meetings

​Workspace builder

​Installation

​Configuration

​Available Instructions

​Teams & channels

​Channel messages

​Chats

​Meetings

​Bot bridge

​DSUL Examples

​Chat connector

​1. Register an Azure Bot

​2. Answer incoming messages

​3. Hand the message to an agent

​What the connector guarantees

​Limits

​Error Handling

​Common Issues

​External Resources

Microsoft Graph API

Tool Agents

Who is this for?

Prerequisites (Microsoft side)

1. Configure the connector

2. Declare the capability in AI Governance

Agent builder

Option A: Enable the shared capability from the catalog

Option B: Run it from your own workspace (recommended)

Brief the agent in its system prompt

Available Tools

Output Formats

Knowledge Resources

Tool Details

teams

channelMessages

chats

meetings

Workspace builder

Installation

Configuration

Available Instructions

Teams & channels

Channel messages

Chats

Meetings

Bot bridge

DSUL Examples

Chat connector

1. Register an Azure Bot

2. Answer incoming messages

3. Hand the message to an agent

What the connector guarantees

Limits

Error Handling

Common Issues

External Resources