Every morning, almost without fail, I spend several minutes each day writing my morning pages. I’m something of an automation evangelist, but my morning pages have nothing to do with efficiency. They’re the opposite of tech-forward—I don’t do them on a tablet or even a computer. Nope. I bust out an old school pen and notebook, and do them by hand . (Cue the gasps from my Gen Z readers).

I’ve spent much of my career coming up with ways to make people’s lives easier; to eradicate dull, tedious tasks that could be better spent doing more interesting work. But morning pages are their own thing. They’re my daily opportunity to clear my mind, organize my thoughts and prepare myself for the day ahead. They can’t be hacked or optimized. And that’s exactly the point.

Organizations are racing to integrate AI into every workflow, with much of the discussion centered on what processes it can take over next. But I'd argue we're still missing a crucial piece of the puzzle—not what should AI do next, but what shouldn’t it do? What work needs to remain distinctly human, not because AI can’t do it, but because something essential will be lost when it does?

The Cost Of Outsourcing It All

AI is already a capable co-worker, and it’s only getting smarter. Powerful new models are being released constantly, handling tasks they wouldn’t have been able to do even a few months ago.

I love that I can now build an AI agent to do everything from organize my email to vibe code an MVP. I also recognize that recent grads may never have to learn foundational skills that grunt work often teaches us. After all, why should a junior analyst spend hours building a financial model from scratch when AI can generate one in minutes? Why should a new hire struggle through drafting a client memo when a polished version is a prompt away?

But some of these shortcuts may be damaging in the long run. According to research from Harvard Business School, AI is increasingly taking over the repetitive tasks that once served as training grounds for developing professional judgment. The risk is that we may inadvertently be creating a whole generation of managers who have never done the work they’re being asked to oversee.

Even seasoned employees are at risk of dulling their hard-won critical thinking skills. MIT Media Lab divided 54 test subjects into three groups and asked them to write SAT essays. The first group had access to ChatGPT, the second had access to Google and the third had no access to any tools at all. After analyzing their brain activity, the researchers found that the group using ChatGPT had the lowest brain engagement of all three groups, and “consistently underperformed at neural, linguistic, and behavioral levels.” The researchers call this "cognitive debt"—the hidden cost of letting AI do our thinking for us.

Where AI-Free Zones Make Sense

None of this is an argument against AI. Instead, it’s recognizing that not all work is created equal, and not all of it is ripe for optimization.

The key distinction is this: some tasks exist for what they produce, and some exist for what they build. AI excels in the first category, and much of what it creates, whether a first draft or block of code, can be done faster than any human. But the second category—work that requires judgment, deepens relationships or sharpens creativity—often requires the struggle that AI is designed to eliminate.

Exactly where these boundaries are set is up to leaders and their teams. The answer may not be immediately obvious, and it may change over time. To help think it through, I like to ask the following questions:

Is this a task where struggling through it builds a skill we value?

Does this work require ethical judgment or nuance that AI can't reliably provide?

Is the relationship at stake—with a client, a colleague, a direct report—one that benefits from a human touch?

This is also part of the reason that our employees are divided into small, cross-functional teams. Not only does everyone have the benefit of learning from one another’s expertise, they’re also actively stretching their teamwork muscles every day. Different perspectives are discussed and debated, assumptions are challenged and decisions are made collectively. In a team setting, AI can make certain tasks easier, but it can’t replace the emotional intelligence that working together develops.

AI is an ultra-powerful tool. But so are our own minds. On its face, writing morning pages may seem like an impractical, inefficient process. That’s okay. I know that they make my thinking sharper, my creativity stronger and my ideas more my own. I also know that I would never have run a successful business for the past two decades by outsourcing my thinking. Some things are worth protecting from optimization—not because AI can't do them, but because doing them ourselves is the whole point.