← Back to all posts

Should AI Do This? Or Just Help You Do It?

by Elena Jäger
Nov 23, 2025
Connect

⏱️ Read time: ~3 min

You try to hand a task over to AI. It doesn't deliver. Now you're wondering if the hype is just… well, hype.

Sound familiar?

I hear this a lot. In my programs, in client work, in sales calls. Coaches and consultants who are frustrated, even disillusioned, because AI didn't do what they thought it would. And when we dig in, the pattern is almost always the same: the wrong tool for the job, combined with unclear instructions.

Let's unpack that: Just because we have powerful AI tools at hand, doesn't mean we should use them for everything.

Why? Let me explain.

Generative AIs, that's your tools like ChatGPT, Claude, Gemini, Copilot, are powered by large language models. In the simplest way, they predict text based on probabilities. They're brilliant at language, ideas, synthesis, and pattern recognition. But they're terrible at tasks that require 100% accuracy.

When you ask AI to do something it's not built for, then instruct it sloppily, the output disappoints. And that's not your fault. It's not AI's fault either.

It's a mismatch between tool and task.

 

Two Examples Where AI Isn't the Answer (But Can Still Help)

 

1. Analyzing large, messy datasets

Feeding raw, uncleaned data into AI (especially non-reasoning models) and expecting flawless analysis? That's a recipe for confusion and errors. The better move: use AI to help you clean the data, spot inconsistencies, suggest fixes, write formulas. Depending on your setup, AI can assist with analysis, identify patterns or even visualize data. But often it's smarter to have it help you with the tasks (draft the code, write the formulas) rather than execute them directly.

 

2. Creating slide presentations

ChatGPT, Claude and their competitors aren't design tools. They can't build you beautiful slides (yet). But they're exceptional at helping you structure your presentation: draft an outline, organize the flow, challenge your logic, think through what content belongs on each slide based on your audience and purpose. Then, AI can format that content so it's ready to copy into a specialized tool like Gamma.

 

The Real Mindset Shift

 

AI-first means leading with the question: "How can AI help me?" not "Can AI do this for me?"

 

Sometimes the answer is:

  • "AI, write me the Excel formula" instead of "AI, be the formula."
  • "AI, draft this code" instead of "AI, execute this logic."
 

Let AI augment your work. Don't expect it to replace your domain knowledge, sharp thinking, and validation. You're not outsourcing judgment. You're amplifying capability.

 

Key Takeaway:

 

Next time you reach for AI, pause and ask yourself: Is this a language or idea task? Or is this a logic or data task? How can AI help me—even if it's not doing the final work?

 

Then use this simple prompt with your AI of choice:

"I would like to (explain task) in order to (explain goal) for (explain target audience). How can you help me?"

 
 

Try this and see if it changes anything.

Have fun experimenting

Elena

Keep Your Expertise in Use When Working with AI
⏱️ Read time: ~4 min I read one section of a leadership development proposal aloud and asked myself: is this how I would actually say it? I had used my AI-powered content library, built from the last seven years of my own work: thought leadership, proposals and actual deliverables. I had done the homework to give AI rich context and clear direction. I also used a slide-generation tool and asked...
I’m Back in Second Grade. On Purpose.
⏱️ Read time: ~2 min I am back in second grade. Well, not literally. But that is the mindset I am choosing as I go deeper into building with AI. I use AI every day. Not just to polish a paragraph, summarize a meeting, or create a quick plan. I code with it. I build applications with it. I connect tools, shape workflows, test ideas, and turn rough concepts into things that actually work. Supe...
Are You in Charge of Your Attention?
⏱️ Read time: ~4 min A few days ago, I was cycling through one of Zurich’s busiest transport hubs when an e-scooter rider nearly hit me. He was moving quickly, faster than the permitted 25 km/h, with one hand on the handlebar and the other holding his phone. His eyes were on the screen, not on the road, not on the cyclist directly in front of him. I swerved. Luckily there was no car beside me. ...

Not signed up yet?
Do it right here:

© 2026 Future of Work
Privacy Policy Imprint Home

JOIN THE VIP LIST

Name of Free Resource

Get started today before this once in a lifetime opportunity expires. Get started today before this once in a lifetime opportunity expires.