Published Date:
April 3, 2025

Start With Tasks, Not Tools

Many ask, “What’s the best AI tool for X?” But the real value of working with AI comes from identifying the right tasks first. This article shows how to make smarter AI decisions that actually support your work.
Daan van Rossum
By
Daan van Rossum
Founder & CEO, FlexOS

The Tool is Not the Point

I ran a workshop this week with a group of about 150 teachers. As often happens, the conversation quickly turned to tools:

“What’s the best AI tool for writing?”

“Should we be using ChatGPT or Microsoft Copilot?”

“Is Gamma better than Beautiful?”

And while these are totally valid questions, I found they’re the wrong place to start.

Because the tool is not the point, the task is.

AI Starts with G.E.D.

As I ​wrote last year​, every exploration into AI should start with G.E.D.

If work is:

  • General (not hyper-niche or reliant on domain secrets),
  • Error-friendly (where it’s okay if the AI makes a mistake or needs review),
  • Digital (the work happens in a digital environment and can be accessed by software).

Then AI should be a partner, or take over completely.

Over the past months, we’ve added two more criteria:

  • Recurring (it’s work we frequently do)
  • Toil (we don’t even like this work to begin with - inspired by ​Debbie Lovich​)

Now, every personal or business-focused AI implementation starts by prioritizing those tasks and workflows that are GED-RT. (AI in HR Today author Antony Onesto has a helpful model for finding those tasks.)

In our ​course​, we go through a simple exercise where we score the parts of our jobs that we spend most of our time on.

  • Is it a G.E.D. task?
  • Do we do it often?
  • Does it take quite a bit of time?
  • Do we not get anything out of it?

If so, AI it.

A recent conversation with Organizational Development expert ​Gwen Stirling Wilkie​ made us move up this prioritization model to the first week of our program, as it is the single most important exercise for busy leaders.

Followed by the tools to solve those GED-RT tasks.

That’s the sequence: tasks first, then the tools. (PS: I prefer ChatGPT and Gamma.)

The Right Tools for You

It can be exciting, sometimes even intoxicating, to be at the bleeding edge of AI tools.

Our ​community​ often buzzes with new platform releases, tool updates, and launches.

But the truth is that:

  1. You can get almost all of it done in ChatGPT or one of the other ​AI models​.
  2. Many tools and apps, even those in the ​AI Top 100​, will eventually be features in bigger platforms.

There’s probably no better example than the ChatGPT image generation update, which (​way beyond Ghibli images​) puts a full-blown designer in your AI team.

While specialist image generator tools like Midjourney and Firefly will have their place, ChatGPT will do the job just fine for most.

AI professor Ethan Mollick ​put it well​ yesterday:

“The fast evolution of AIs makes it risky to spend tons of effort getting around current model limits through clever approaches, rather than waiting.

Rather than trying to stay current on everything, it is much better to start with your use case where AI would actually make a meaningful difference and then find the right platform for it.

Plus, it will keep you sane versus drinking from the AI firehose.

From Doing to Deciding: The Changing Role of Humans

In the same workshop, I made a point that deeply resonated, especially as these educators prepare the next generation to enter the world of work.

The acceleration of AI capabilities is insane.

I don’t see another reasonable word for it.

We’ve gone from zero AI in most workflows just three years ago, via the launch of ChatGPT’s text-based interface, to a moment where AI can talk, listen, see, and create—in real time.

For most ​SuperWorkers​, those who’ve already reinvented their roles for the AI age, it’s a daily habit to have a voice call with an AI advisor, let it turn meeting transcripts into proposals on autopilot, and have AI agents doing their research.

But this video, a year old, still shocks most people.

Can AI really do that?

Yes, it can.

And this is where the next step is for most.

After finding the right tools for your tasks, think about which tasks you’ll even be doing in the future.

If we extrapolate from what has happened in just 2.5 years, what does your job, your industry, look like in just a few more years?

What will you do? What do you WANT to do?

​Antony Slumbers​, one of my favorite voices in this space, ​once told me​ that in a world where technology is handling more and more, we have to become better at being human.

That means focusing on work's messy, unstructured, interpersonal, and creative parts. Not competing with machines on speed or output:

“We need to lean into the work that is non-structured, non-repeatable, and non-predictable. AI excels at repeatability and predictability—but we excel at ambiguity, at care, at imagination.” – Antony Slumbers

It’s important to note that in most knowledge work jobs, what’s actually being replaced is not the whole job—it’s parts of the task bundle.

A financial analyst might no longer build the model, but instead spend time telling the story behind the data. A marketer might not write the campaign copy, but they’ll decide the brand voice and emotional tone.

This means the high-value human contributions now center around:

  • Judgment and ethics
  • Creativity and originality
  • Empathy and leadership
  • Problem framing and decision-making
  • Big-picture thinking

In a world where AI can do most of what we do today, what matters is taste, curation, and creativity. So while AI may do the work, humans will decide what’s worth doing.

So what will you do?

Let’s not stop at doing the same, but faster.

Expand your capabilities. Increase the quality of your work. And fundamentally rethink what ‘work’ means.

LEAD WITH AI

🔥 Become the Most AI-Savvy Leader in the Room

Our ​Lead with AI Executive Bootcamp​ is now Cohort 8 - And it's our most powerful version yet!

We’ve rebuilt the syllabus from the ground up to cover the breakthroughs everyone’s talking about, but in ways that you can actually apply to your executive workflows:

  • Your AI Video Avatar that can scale your thought leadership videos
  • "Vibe Coding" lets you make an app or site in minutes just by talking
  • Leverage Multimodal AIs to their best (incl. the new ChatGPT-4o Image Generator and Advanced Voice Mode)

And beyond the breakthroughs, you’ll get the proven lessons loved by 250+ senior leaders, designed to help you master advanced prompting, build custom GPTs, automate tasks, and lead smarter with AI.

​View the full syllabus and new lessons here.​

Sign Up Today and Master AI in Just 3 Weeks - $200 OFF Ends Soon!