An AI Guide for Managers: Rebuilding the Decision, Report and Meeting Flow (2026)

For managers, the value of AI lies not in learning the tools but in rebuilding the decision and report flow. A manager's week is spent preparing for meetings, reading reports, on email, giving team feedback and writing decision rationales. Most of this work is text-based, and that is exactly where AI is useful. This guide explains, from a manager's point of view, which task is moved how, where to stop and how the team's use is managed.

The article follows the flow of the workshops I run with manager groups. No technical knowledge is needed. For the format and content, see the AI training for managers page.

Four flows the manager uses personally

1. Decision preparation

The documents on the matter to be decided (proposal, report, previous correspondence) are given to the tool through a corporate account, and three things are requested: the options and the risk of each, unanswered questions, missing data. The tool does not decide. It organises how the decision comes to the table. This single step largely eliminates the sentence "let's look into this again" in meetings.

2. The weekly report

Scattered updates from the team, spreadsheet outputs and notes are converted into a single "weekly status" format: what was finished, what is late, which decision is pending. The report format is set up once and produced in the same pattern every week. The version for senior management and the version for the team come out with a single request. For a detailed setup, see the reporting automation guide.

3. Before and after meetings

Beforehand, three questions and two possible decisions are requested from the agenda and previous notes. Afterwards, the meeting notes are given to the tool and a table of decisions, owners and dates is produced. Tools that capture meeting summaries automatically (the assistants inside Teams and Meet) make this easier, but the accuracy of the summary is always checked.

4. Difficult emails and feedback

For bad news, a rejected request or performance feedback, the text is first written freely, then the tool is asked for a tone correction and two alternatives. The text the tool writes is never sent directly. It is translated into the manager's own voice.

Tasks delegated to the team, and the manager's rule

Tasks at team level vary by department: sales scripts, marketing content, reconciliation checks, contract summaries. The manager's job is not to do these tasks but to set three rules:

Without these three rules, team training reverts to chat-window use within a week. With the rules, the training sticks. For setting up strategy at company level, the article on how to build an AI strategy describes the next step up.

Where to stop: the manager's three red lines

• Personnel decisions: For hiring, promotion, exit and pay decisions the tool only prepares. The decision rationale is written in the manager's own sentences.
• Numerical accuracy: The tool summarises the table but does not verify the figure. Before a financial figure goes into a report it is compared with its source.
• External communication: Text that goes to a customer, the press or a partner is not sent as it came out of the tool.

A representative scenario: the general manager of a 30-person company

The scenario is representative, not a measured client result. Suppose the general manager runs six meetings a week, receives reports from three departments and answers forty emails a day. After the training:

• Before every meeting a ten-minute preparation flow is run, and meetings get shorter and end with a decision.
• The three departments' reports arrive in a single format, because the format came from the manager.
• For difficult emails the habit of "I write first, then have it corrected" is established.
• A one-page data rule is published for the team, and at the end of the first month which tasks were tied to a flow is reviewed.

The gain in this scenario is not speed. The gain is that the manager's day shifts from gathering information to making decisions.

Which tool?

Tool selection for a manager differs from the team's: the best tool is the one that talks to the office suite the company already uses. In a company on Microsoft 365, Copilot is the natural start. In a company on Google Workspace, Gemini and NotebookLM. Alongside these, Claude stays on the table for long document analysis and writing quality, and ChatGPT for general-purpose use. The number of tools is kept small, because for a manager the issue is not the tool but the flow.

How to start

A half-day workshop is enough for a manager group: everyone works with their own meeting notes, their own weekly report and their own difficult email, and leaves at the end of the day with a personal flow and a team rule draft. For the content, see AI training for managers. For company-wide strategy and pilot projects, see the AI consulting page. If you want to try it yourself first, a one-to-one session is the fastest route.