Microsoft 365 Copilot: turn email into a table

Key insights

• The example uses Copilot Chat to review request emails in a shared mailbox.
• A useful prompt states the goal, names the source, and describes the desired answer.
• Check the response against its sources, especially dates Copilot had to infer.
• A follow-up prompt can refine the result before you take it into Excel.
• The example requires a Microsoft 365 Copilot license.

Microsoft 365 Copilot Chat can turn varied requests in a shared mailbox into a structured table, giving users a consolidated view instead of leaving details scattered across messages. This approach is useful for people who handle incoming requests and need to organize the information for review or follow-up. Damien Bird demonstrates the workflow with sample emails for a fictional distillery.

What it does

The task is to make sense of messages that describe requests in different ways and bring their details together in one table. Rather than copying information from each email by hand, a user asks Copilot Chat to work from the mailbox and present an answer in a defined format. The example focuses on cask and event requests arriving in a shared inbox.

The prompt has three parts: say the goal, name the source, and describe the answer. These instructions give Copilot a task, identify the information to use, and specify the form of the response. The distinction matters: asking for a general summary is not the same as asking for information organized into a table.

A table can make a collection of requests easier to scan, but the output still needs checking. In the example, Copilot provides sources for its answer, and some dates require inference. That is a reason to verify extracted details against the original messages rather than treating the table as a confirmed record.

Who it is for and what you need

This workflow is for Microsoft 365 Copilot users who need to process email requests, particularly people working from a shared mailbox. The stated requirement is a Microsoft 365 Copilot license. The example begins in Copilot Chat with access to the shared mailbox being used for the task.

The emails and distillery in the demonstration are fictional, and the messages are sample data. That makes the example a way to understand the prompting pattern, not a real operational record. In your own work, choose an appropriate set of messages and make sure you can access the mailbox information you want Copilot to use.

The key user decision is what the finished table should contain. Decide which details will help you handle the requests, then express the desired result in your prompt. Naming the source and the required format helps keep the task focused on the information you intend to organize.

How it works

1. Open Copilot Chat and identify the shared mailbox or email source you want it to review.
2. State the goal: explain that you want the requests collected or organized, rather than asking only for a broad summary.
3. Name the source so Copilot knows which emails to use.
4. Describe the answer you want, including that it should be a table and the details it should capture.
5. Review Copilot’s response and check its cited sources against the underlying emails. Pay close attention to dates that Copilot inferred.
6. Use a follow-up prompt to refine the result if needed, then take the table into Excel for further work.

The sequence is intentionally iterative. The first answer gives you something to inspect; the source check helps expose details that deserve attention; and a follow-up lets you clarify the requested output. Moving the result into Excel comes after that review, rather than replacing it.

Copilot Chat can also be used to try the Researcher agent for a sourced report. The demonstration describes typing @ to add an agent to a prompt. This is a separate option from turning the email requests into a table, but it illustrates that the prompt can include an agent when the task calls for one.

Limitations and alternatives

Copilot’s answer should be treated as a draft to verify. The example calls attention to dates that had to be inferred, so check those against the original emails before using them to plan or track work. More generally, confirm the table’s contents against its source messages before relying on it.

How useful the result is depends in part on how clearly you describe the goal, source, and desired answer. If the output needs adjusting, refine it with a follow-up prompt. Taking the result into Excel is the next step shown when you want to work with the table there; the email review and the spreadsheet work are distinct parts of the process.

The scenario uses sample email data, so it does not establish how every shared-mailbox request will be interpreted. Start with a clear request, inspect the cited source material, and check any inferred details before acting on them. This keeps the user responsible for validating the information Copilot organizes.

Sources

• Damien Bird’s Microsoft 365 Copilot demonstration

Frequently asked questions

What license do I need to organize emails with Copilot Chat?

The workflow specifies a Microsoft 365 Copilot license.

How do I prompt Copilot to turn email requests into a table?

State the goal, name the email source, and describe the answer format you want, including a table.

Should I check Copilot’s email table before using it?

Yes. Check the answer against its sources, especially dates Copilot inferred, and refine the response if needed.