*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.
AI tools are currently evolving so quickly that even my own use of them is constantly changing. Features that weren’t available a few months ago are now part of my daily work routine. At the same time, I’m realizing more and more that for me, it’s no longer just about asking ChatGPT, Claude, or Gemini a question and getting an answer.
AI is increasingly becoming part of my processes. I connect them to other systems, have documents created, prepare emails, review meeting transcripts, plan trips, and even incorporate language learning into my regular workday.
In this post, I’d like to give an update on how I’m actually using AI right now—using various examples from my everyday life. The examples are taken from several videos from the past few months and thus clearly illustrate how my use of the platform has evolved step by step.
ChatGPT Work and Claude: Using Two AI Tools Simultaneously
Over the past few weeks, I’ve been working on many tasks simultaneously using ChatGPT Work and Claude. Time and again, I find that these tools each have their own strengths.
I find ChatGPT Work particularly exciting when I’m working on more complex tasks and want to integrate different systems. For example, plugins can be used to integrate email systems such as Outlook or Gmail, as well as other applications. You can also enter your own skills, such as those related to your personal writing style.
On the other hand, I particularly like Claude for certain tasks because of the depth of his answers. That’s why I no longer base my decisions on the question, “Which is the best AI tool?” but rather on which tool is best suited for my current task.
Another important point: Office documents are increasingly being created directly within AI processes. Word documents, Excel files, and PowerPoint presentations can not only be generated based on their content, but can now also be formatted in such a way that they can actually be reused later.
For me, this development is what really matters: AI no longer just generates text, but increasingly delivers a finished product.
Don’t just write emails—save them directly as drafts
One example that particularly illustrates how I currently use AI is email communication.
I use Claude in such a way that all I have to do is describe what I want to write to whom. I have already specified my preferred key and other requirements. The AI then drafts the message and can save the draft directly to my email program.
It gets particularly interesting when replying to existing emails. For example, I can say something along these lines:
“Reply to the last email from XY and write that …”
The AI searches for the relevant message, generates a suggested response, and uses it to create a draft. In one of my examples, the system even detected that the original message came from a “no-reply” address. Instead, another contact address from the footer was suggested.
Of course, I check the results before shipping. But it’s precisely this last point that’s important to me: I no longer have to switch between different applications, copy text, and paste it again. The AI will be integrated directly into my existing process.
Use meeting transcripts as a starting point for further work
Meeting transcripts are also among the things I now regularly reuse. That said, I almost think it’s a shame to use a transcript just to create a standard meeting minutes.
This is because a transcript contains a great deal of context that can be used for other tasks.
For example, I use meeting transcripts to:
- Summarizing meetings in a structured way,
- To identify tasks and responsibilities,
- to create a presentation outline based on a conversation,
- To analyze risks, unstated assumptions, and potential opportunities,
- Prepare follow-up messages for the participants.
I find the analysis of what lies between the lines particularly interesting. For example, if after a meeting you’re left with the feeling that something didn’t go quite right, AI can help bring potential implicit problems or assumptions to light.
Of course, the same applies here: AI results must be verified. You should be especially careful with prompts that ask the AI to fill in missing information on its own. This can quickly lead to assumptions that weren’t even part of the original meeting.
Data protection remains a critical issue
The more we integrate AI into our existing systems, the more important data protection becomes.
This applies to access to emails and company data as well as meeting transcripts. If you record or transcribe your meetings, participants must be informed of this. At the same time, you should carefully review which tool you use for which data.
I therefore make a conscious distinction between applications where data protection is particularly important and tasks that involve the processing of less sensitive information.
Direct system integration offers a great deal of convenience. But that is precisely why we should be mindful of what data we make available to an AI system.
Create a Website in Just a Few Minutes with AI
Another field that is growing rapidly is website development.
For example, I tried out ChatGPT Sites and had it create a page about me based on a very simple prompt. The goal was to showcase my professional experience and attract potential new clients.
An initial website design was created in a short period of time. In the process, content and images from my existing website were incorporated, partially restructured, and supplemented with new text.
I find this approach exciting for a quick sketch or a last-minute page for a specific occasion.
However, when it comes to creating a productive website, I currently prefer a different approach: For example, I use Claude to generate an HTML file through a conversation. A ” index.html ” created in this way can then be uploaded to your own hosting provider and published under your own domain. That way, I was able to create a complete website—including clickable elements and a language selector—in just a few hours.
To me, this really illustrates how much AI can lower the barrier to entry for technical tasks.
Combining Learning and Work
One of my favorite apps right now actually has nothing to do with productivity in the traditional sense: I use Claude both as a work assistant and an Italian tutor.
To that end, I’ve set up a separate project with a more detailed system prompt. The basic idea is simple:
I write my inquiries in Italian. First, Claude corrects my Italian and explains my mistakes to me. The AI then responds in Italian. However, the actual work products—such as presentations or documents—will continue to be created in German.
For example, I can ask in Italian to create a presentation on a specific topic. My input is corrected, communication takes place in Italian, and yet I still receive the finished presentation in German.
That way, I incorporate learning into something I do every day anyway.
To me, that’s a fascinating idea—one that, of course, isn’t limited to Italian. You can use the same method for other languages or even other learning goals.
Control AI with your voice instead of typing everything out
Another small step with a surprisingly big impact is voice input.
Instead of typing out every prompt, I’ve started talking to my AI assistants more often. I prefer a combination of both: I give my instructions verbally, but still have the results displayed in writing or visually.
This is especially convenient for longer prompts. I can explain what I want with relative freedom, add details as needed, and follow my train of thought without having to come up with a perfect prompt beforehand.
This also shows that a good prompt doesn’t always have to be perfectly written. Typos or slightly disjointed phrasing are no longer a major problem for many modern AI models. It’s much more important that you provide enough relevant context and clear requirements.
Planning Bike Rides and Trips with AI
AI now helps me with my personal planning tasks as well.
For a bike tour around Lake Como, for example, I gave Claude various requirements: traveling by train with a bike, a specific route length, as few hills as possible, no unpaved surfaces, and several route suggestions.
After receiving an initial response, I further specified my requirements and designated a specific departure station. Afterward, I had the information compiled into a Word document. I found it especially convenient that it included links to Google Maps for the suggested routes.
To me, this is a great example of working with AI: the first prompt doesn’t have to be perfect just yet.
You can start with a rough idea, take a look at the result, and then use the dialogue to specify what’s missing or what needs to be changed.
From Individual Prompts to Real AI Workflows
When I look back on my use of AI over the past few months, one trend stands out: I’m using AI less and less as a standalone chatbot.
Instead, small workflows are created.
A meeting is transcribed, and that transcript is used to create tasks, a presentation, and a follow-up message. From just a few bullet points, an email is created and saved directly as a draft in your inbox. An idea for a website becomes an HTML file through a collaborative process. A travel idea is turned into specific route suggestions and a structured document.
And even my Italian studies have now become part of my normal work routine.
That, to me, is exactly where the great value of AI lies right now: Not every single task has to be spectacular. What really matters is how well AI can be integrated into the many small routines of our daily lives.
There are two things you shouldn’t forget: Check the results and keep data protection in mind. If these fundamentals are in place, it’s worth experimenting regularly and rethinking existing processes.
So perhaps the more exciting question is no longer just: “What can AI do?”
Instead: “At what point in my daily routine can AI already take a task off my hands?”
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.
