One AI Agent, Three Apps: A Morning Brief From Outlook, Calendar, and Slack

Most morning briefs begin with three separate searches.

What arrived in email? What is on the calendar? What happened in Slack while you were away?

Each question is simple. The work comes from opening every app, separating decisions from noise, and figuring out what deserves attention first. By the time the brief is finished, a meaningful part of the morning is already gone.

That is exactly the kind of work we want an AI assistant to handle.

Instead of checking three apps, imagine starting with one request:

Give me a founder morning brief. Review unread Outlook email since yesterday, today’s calendar, and Slack mentions. Return only the three decisions I need to make, the three biggest risks, and the next action for each. Draft any useful replies, but ask before sending or posting anything.

This sounds like one prompt. In practice, it asks the assistant to complete a real workflow.

From three feeds to one decision queue

A useful morning brief is not a longer notification summary. It should reduce the amount of information a person has to process.

With configured Microsoft 365 and Slack connections, Pneum.ai can bring the relevant context into one conversation. For a morning brief, that means it can:

1. Review the allowed Outlook messages, calendar events, and Slack activity.
2. Separate decisions and risks from routine updates.
3. Connect related information across services.
4. Suggest a clear next action for each important item.
5. Draft useful replies without sending them automatically.

The result is a decision queue: what needs your judgment, why it matters, and what could happen next.

The video above shows a smaller version of this interaction pattern. Pneum.ai reads calendar and file context through Microsoft 365, presents a proposed change for approval, and reports back after the work is complete.

The approval moment is the point

The most important part of the request is the final sentence: “Ask before sending or posting anything.”

Reading a calendar and drafting a response are different from changing an event or sending a message to another person. Pneum.ai is designed to make that boundary visible.

When an action requires approval under your configured policy, the request appears in the conversation before the action happens. You can review what will change, approve it, or decline it. Permissions can also be scoped by profile, tool, connected service, and user instead of giving one assistant permanent access to everything.

That keeps the workflow fast without removing the person responsible for the result.

One conversation can coordinate several specialists

Requests that cross multiple services can involve many searches and tool calls. Making the main assistant do all of them in sequence would turn a quick conversation into a long wait.

Pneum.ai can delegate focused parts of the work to specialist agents while keeping the main conversation responsive. Their activity remains visible, and you can stop or redirect work when needed.

For the morning brief, one specialist might gather communication context while another checks the schedule. Pneum.ai can then bring their work back into one answer instead of leaving you to combine several outputs yourself.

This same pattern extends beyond a daily brief:

• Compare an Outlook sales thread with the latest proposal in OneDrive.
• Turn a long Slack launch discussion into decisions, blockers, and owners.
• Find open meeting times and prepare a reply without sending it.
• Compare a SharePoint document with recent team discussion and flag disagreements.

The user asks for the outcome. The assistant coordinates the work.

Connected work without a workflow builder

Traditional automation is powerful, but it often begins with triggers, branches, field mappings, and a workflow that has to be maintained when the process changes.

Conversation is a more natural starting point for work that changes from day to day. You can ask for three decisions today, focus only on customer risks tomorrow, or add another source without rebuilding a diagram.

Pneum.ai now supports connected services including Microsoft 365, Slack, GitHub, Jira, Salesforce, Philips Hue, Telegram, and Discord. Availability and allowed actions depend on the service, account, and permissions you choose.

The goal is not unlimited automation. It is one understandable place to ask, review progress, approve important actions, and receive the result.

Try the interaction before connecting anything

You can try a focused version of Pneum.ai in the free live demo. Speak or type to the assistant, ask it to work with temporary notes and reminders, and see how approval requests appear before a change is made.

The public demo is intentionally isolated. It does not connect to your Microsoft 365 or Slack account, and its temporary conversation state is cleared when the session ends. The recorded workflow above uses a separately configured Microsoft 365 connection.

The demo is enough to experience the part that matters: describe an outcome naturally, watch the assistant work, and stay in control of what happens next.

A morning brief should not take all morning. It should give you the decisions, risks, and next actions you need, then let you get on with the day.