Why knowing multiple AI tools matters (and how they differ)

When I was preparing my talk for AI Predictions 2025 – showcasing the most important AI trends, I decided to run a small experiment.

I asked participants from my latest AI Leadership Masterclass what trends they expect for 2025. Then I fed their answers into four different AI tools:

  • Claude
  • ChatGPT
  • Google NotebookLM
  • Microsoft Copilot (yes, it runs on GPT, but it behaves a bit differently)

Finally, I compared the results.

What I discovered

First, the good news: the participants see exactly how AI will reshape their work and business. More interesting, though, is how differently each tool handles the same data:

  • Claude – a more business-oriented angle, human-sounding prose, and the ability to code micro-apps right in the browser.
  • ChatGPT – great at web browsing, plugging into add-on tools (GPTs), data analysis, and even image generation.
  • Google NotebookLM – excels at huge data sets (PDFs, videos) and always cites sources for fact-checking.
  • Microsoft Copilot – tightly integrated with Office apps; perfect if you already live in the Microsoft ecosystem.

And that’s not all. For working with large slide decks, I found Google Gemini the most useful (it handles longer texts than Copilot). I then refined those outputs in Claude, because its analyses and wording strike me as clearer and more creative.

Every AI tool has its own talents – and tone

This experiment confirmed two key points:

1. Each tool has a unique “skill set.”

Some are better for creative writing, others for data crunching, others for corporate app integration. No single universal tool exists – yet.

2. Each tool has its own voice.

The analyses revealed distinct “personalities”:

  • Claude offers a more businesslens perspective.
  • Copilot drills into technical details.
  • ChatGPT is bold and ambitious in its predictions.
  • NotebookLM pulls in broader societal context.

Paul Skallas nailed it in a tweet:

“Claude is impressive. You can really see the craftsmanship. But it has a personality that’s a bit overpowering. ChatGPT is dumber and simpler…but you can steer it. It doesn’t steer you. It’s like driving a ’90s car versus a modern one.”

What this means for companies

Choosing an AI tool will shape not only your workflow but the direction of your entire organization. Because each model thinks and communicates differently, it influences how people inside the company think and decide. My recommendations:

  1. Keep several AI tools handy – pin Claude, ChatGPT, Copilot, and others to your browser.
  2. Compare outputs – run the same task in multiple models and observe the differences.
  3. Discuss as a team – share what works best for whom, and why.
  4. Prompt smartly – you don’t have to craft the perfect prompt yourself; let AI help you improve it.

A prompt for better prompts

Try this simple trick:

  1. Start with: “Write instructions for an AI agent that will…”
  2. Add the task (e.g., process meeting transcripts, draft emails, create slide content).
  3. Include a few specifics or constraints.
  4. Finish with: “Based on your experience with similar tasks, suggest how this instruction could be improved.”

You’ll get a prompt better than one you’d write on your own.

In short

Working with AI isn’t about finding one perfect tool. It’s about experimenting, noticing the differences, and building a workflow that suits you. And who better to help than AI itself?

So, will you give it a try? Which tools work best for you? Share this piece with your colleagues. Tag me if you do; I won’t mind one bit. 😉

F.D.