ChatGPT for Outlook: Draft Emails and Track Follow-Ups

ChatGPT can help you draft, revise, and respond to emails in Outlook — even without a direct integration. The most dependable approach is bringing email content to ChatGPT in a browser tab: copy the thread, paste it in, and write a clear instruction. Tools like fenbs keep the prompts that work well and the resulting follow-up tasks in one place, so you do not rewrite from scratch each time and nothing falls through after the session ends.

For Microsoft 365 subscribers, Copilot in Outlook offers the same kind of drafting assistance without leaving the email interface. These are different products that serve similar purposes, and which one fits your workflow depends partly on what your organization already licenses.

Your options for using ChatGPT with Outlook

There are three main ways to use ChatGPT alongside Outlook:

• Copy and paste: open the email, select the text, and paste it into ChatGPT at chat.openai.com. Add your instruction and review the draft before using it. This gives you full control over what you share.
• Third-party add-ins: tools published in the Microsoft AppSource marketplace can connect ChatGPT to Outlook. They vary in quality and data handling. See the companion article on the ChatGPT for Outlook add-in for what to check before installing.
• Automation workflows: Power Automate and tools like Zapier can connect Outlook to the OpenAI API for higher-volume tasks. These require setup and careful permission configuration.

For most individual users, copy and paste is the fastest start. The drafts come back as text you review and edit before sending. Nothing is sent automatically, which matters when a small wording change can make a significant difference to the reader.

How to write email prompts that work

A prompt that reliably produces a good email draft has three parts: who you are writing to, what you need to communicate, and the tone you want. "Write a reply to a client asking to reschedule our Friday call to next week, keeping a professional but friendly tone" gives ChatGPT enough to produce something usable.

Including the original email thread helps ChatGPT match the existing conversation and avoid repeating something that was already said. Paste the thread above your instruction rather than below it: ChatGPT reads from the beginning, so context first, instruction second.

Always read the draft before sending. Watch for details that are technically accurate but wrong in context — a friendly tone that reads as too casual for the situation, or a summary that leaves out the most important request. ChatGPT drafts are a starting point, not a finished product.

Checking AI-generated replies before you send

Email carries your name and reputation. A generated reply that misrepresents a commitment, overpromises a deadline, or misreads a tone can create real problems. Read the draft as if you received it: does it say what you mean? Does it answer the question that was actually asked?

It helps to check the key facts explicitly. If the draft mentions a date, verify it is correct. If it references a previous conversation, make sure the reference is accurate. ChatGPT can confidently state something that is wrong if the email thread gave it incomplete information.

Handling sensitive or confidential email

Not every email should go to ChatGPT. Emails that contain personal information about clients, legally privileged communications, or confidential business details deserve more caution. Check your organization's data handling policy and your company's AI use guidelines before pasting sensitive content into any external AI service.

For sensitive drafts, an alternative is to paste in a sanitized version: remove names, identifying details, and specific figures, replace them with placeholders, and ask ChatGPT to help with the structure and tone. You then fill in the specifics yourself before sending. This gives you the drafting help without sharing the details.

Saving the prompts that work

When you find a prompt that consistently produces good drafts for a particular type of email — a client update, a vendor follow-up, a meeting confirmation — save it. You lose a surprisingly large amount of time to rewriting prompts from memory when they could be a one-second paste.

fenbs has a Skills and prompts library where you can save prompts alongside the project they belong to. Any connected AI assistant can read the saved prompt and use it with the current project context, which means you do not have to rebuild the instruction each time. See how it works for the full prompts feature.

How to do this with fenbs

fenbs does not connect to Outlook directly. It holds the prompts that work and the follow-up tasks your email creates, so neither gets lost after the session.

1. Create a free account. Add a project for the client or workflow where email drafting matters.
2. Save each email prompt that works as a Skill in the fenbs Skills and prompts library: the instruction, the tone note, any client-specific rules. Add it to the project it belongs to.
3. Use ChatGPT to draft the email. Paste the thread and your saved prompt. Review the draft carefully before sending.
4. When an email creates a follow-up task, add it to fenbs with a note about the source. "Promised revised proposal by November 1 — see October 21 email."
5. Connect a supported AI app from Integrations. At the start of the next session, ask it to read the project context and the saved prompts before working on new email.