*Thisarticle is based on a YouTube video by Sophie Hundertmark, an expert in the use of artificial intelligence with a focus on chatbots and strategic AI applications in companies and public institutions. Sophie is a researcher and lecturer at the Lucerne University of Applied Sciences and Arts and is doing her doctorate in Conversational AI at the University of Fribourg. The blog text was created using a custom GPT model that was trained on Sophie’s video content, language style and expertise. The result is well-founded, up-to-date articles based on Sophie Hundertmark‘s own expertise.
You can find the link to the video at the end of this article.
Meetings are over in a flash, but the real work often doesn’t begin until afterward. Decisions must be documented, next steps must be recorded, and follow-up messages must be sent. Especially when there are a few days between the meeting and the follow-up, it’s not so easy to remember all the important details.
This is exactly where AI tools can help. With a meeting transcript and the right prompt, you can have a structured follow-up message generated in no time. The AI not only summarizes the conversation, but can also identify key decisions and next steps from the conversation and use them to draft a message directly.
However, it is important to note that before a meeting is recorded, transcribed, and then processed using AI, data protection, participant consent, and—where applicable—internal guidelines must be taken into account.
From Meeting Transcript to Finished Follow-Up Message
The basis for this use case is a transcript of the meeting. Many meeting and video conferencing tools now offer the option to automatically transcribe conversations.
You can then feed this transcript into a suitable AI tool. Instead of simply asking the AI for a summary, it’s worth defining more precisely what kind of result you need.
A possible prompt for this is:
Create a follow-up message based on the following meeting transcript.
Sound: clear, friendly, professional.
Content: brief summary, key decisions, next steps.
Write as if the message came directly from me.Transcript:
[Insert meeting transcript]
This gives the AI a clear task and, at the same time, lets it know how the result should be structured and phrased.
Why a clearly structured prompt yields better results
When working with AI, the following applies: The more clearly you describe what you need, the more effectively the tool can assist you.
In this example, the prompt contains four key pieces of information:
- Task: Write a follow-up message based on a meeting transcript.
- Tone: The message should be clear, friendly, and professional.
- Contents: The summary, decisions, and next steps should be taken into account.
- Perspective: The text should sound as if you wrote it yourself.
This structure can easily be adapted to your own needs. For example, you can specify that the message should be particularly short, that certain people should be addressed directly, or that the next steps should be presented as a list.
Here’s How to Bring Your Own Personal Tone to AI
In many cases, the instruction “Write as if the message were coming from me” alone is not enough. After all, the AI first needs to know how you usually write.
To do this, you can, for example, provide the AI tool with a few of your own emails or follow-up messages as examples. These texts provide the AI with additional context about your communication: Do you tend to write briefly and directly, or in more detail? How do you start and end your messages? What phrases do you use regularly?
However, the same rule applies here: Use only content that you are permitted to process with the respective AI tool, and be especially careful with personal or confidential information, particularly in the case of real emails.
If you use this use case regularly, you can save your tone of voice and other guidelines as a reusable context. That way, you won’t have to explain all over again in every follow-up message what the result should sound like.
This is especially helpful if the meeting took place a few days ago
The practical benefits of this workflow are particularly evident when follow-up work is not done immediately after a meeting.
Maybe you’ve been in this situation: A week after a meeting, you want to write a follow-up message, and while you still have a general idea of what was discussed—you can’t remember the specifics, what decision was actually made, or who was supposed to take on which task.
Instead of going through the entire transcript yourself again, you can have the AI search specifically for this information. AI tools can analyze lengthy texts, identify relevant content, and use it to create a clear and concise report.
This doesn’t just save time when drafting. It can also help you organize information from the meeting in a more systematic way.
Privacy: Don’t just record and transcribe meetings without asking first
As convenient as this workflow is, you must consider data protection before recording or transcribing a meeting.
Just because a device has a recording or transcription feature doesn’t automatically mean you can easily record every conversation. Inform participants in advance and in a transparent manner if a meeting is to be recorded or transcribed, and obtain the necessary consent.
The subsequent processing by an AI tool should also be taken into account. For example, a meeting transcript may contain names, personal data, customer information, internal strategies, confidential projects, or other sensitive company information.
Before you enter a transcript into an AI tool, you should therefore clarify the following, among other things:
- Was the necessary consent obtained for the recording and transcription?
- What information does the transcript contain?
- Is it permissible to process this information using the selected AI tool?
- What privacy and security policies apply to the tool being used?
- Does your company have internal guidelines for the use of AI?
- Should certain information be removed or anonymized before processing?
The specific requirements that apply depend, among other things, on your company, the context of the meeting, the data being processed, and the tools used.
Therefore, the workflow should follow this order: First, clarify data protection, consent, and internal guidelines—then record, transcribe, and continue working with AI.
The AI creates the draft—you do the final check
Even if you have a good prompt, you should still review the generated follow-up message before sending it. It’s especially worth checking against the original transcript when it comes to decisions, responsibilities, deadlines, or other important details.
For example, pay attention to whether:
- all relevant decisions were accurately reflected,
- Responsibilities are assigned correctly,
- Schedules and deadlines are on track,
- no important next steps are missing,
- no statements were added that were not actually made during the meeting,
- that the wording actually fits your personal communication style.
AI thus handles a large part of the preliminary work. However, the final review and responsibility for the message remain with you.
A simple prompt can become a reusable workflow
This use case becomes particularly interesting when you use it more than just once. The prompt can be saved as a template and reused after every relevant meeting.
Over time, you can continue to refine it and, for example, add your own personal tone, preferred structure, or recurring requirements.
This creates a simple workflow for your daily routine:
Clarify privacy and consent → Conduct the meeting → Create a transcript → Generate an AI draft → Review the results → Send a follow-up
It is precisely these clearly defined tasks that demonstrate how AI can be put to good use in everyday work. It’s not about handing over all communication to an AI tool. Instead, the AI handles time-consuming preparatory work and helps you turn a lengthy meeting transcript into a structured follow-up message more quickly.
At the same time, handling conversation data responsibly should be part of this workflow from the very beginning—and not something that is only addressed once the transcript has already been imported into the AI tool.
Any further questions?
Do you have any questions? I am happy to support you, act as a sparring partner and answer your questions. I am always happy to receive your messages, preferably by WhatsApp message or e-mail.
