PLN 44: The One Thing a Leader Should Never Delegate to AI

Time to Read: 4 mins 

I'm not against AI. Not at all. I use it on the daily for different tasks. 

But I'm seeing more and more instances of leaders outsourcing their judgment along with their tasks, and it's creating real business and leadership challenges.

A colleague of mine, "James," spent the last year automating almost everything he could. AI agents now handle his billing, follow-up, lead generation, and competitive research. 

And it was a move that freed up enough time and bandwidth for him to explore different ways of growing his business. He decided expanding the sales team was the best way to do so.

He also determined it was important to keep the actual sales process human. For context, his company creates custom solutions around what each client specifically needs, so he wasn't willing to hand that piece over to AI.

At least, that's the story he initially told himself.

Fast-forward to when his newest sales rep landed a great opportunity. As part of the training process, they sat down to build the proposal together.

The rep had already met with the prospect and conducted a needs assessment, so James showed him how to feed the info into AI using a prompt he'd been working on perfecting for quite some time. It generated a custom proposal that read well, included their branding, and looked aesthetically pleasing.

With the snazzy proposal in hand, they headed to the client meeting… where everything fell apart.

On the surface, the proposal looked like it understood the prospect's business. But it didn't. There was no real read on their nuanced situation. It felt like "lipstick on a pig," as James shared.

He told me he was embarrassed, because the potential client was right. Their details were included, yet it read like a generic solution.

Essentially, the proposal said a whole lot, without saying anything at all. And, as you might expect, they lost the deal.

But it was even worse than that!

James told me his sales rep later shared how uneasy he felt about the proposal because the AI regurgitation lost the human understanding of the client's needs. BUT, the rep didn't say anything at the time because it was a new job, James is his boss, and his boss was the one training him to do it that way.

So, in one fell swoop, James lost the deal… and his sales rep's trust.

All because judgment got automated along with the task.

But task ownership and judgment ownership are two very different things. And in the world we now live in, we have to pay attention to when AI is blurring the line between the two.

That's on us, and that takes self-leadership.

Because the truth is, AI is very good at producing something that looks like judgment. It sounds confident, uses polished language, and gives the appearance of custom thinking.

But none of this is the same as having actually thought it through ourselves.

My conversation with James has shifted how I personally approach AI. Now, instead of asking "What can I hand off to AI?", I’m asking…

"Which parts should I keep doing myself because they keep my judgment sharp?"

And the wording I'm using for that question is quite intentional, so let's talk about it.

Judgment isn't something you either have or don't. Like a muscle, it's something that gets stronger with continued use.

The parts of your work you stop doing yourself are the parts you slowly stop being able to do well. It’s atrophy in motion.

The key is to catch the atrophy before it costs you something important… like a large sales opportunity, or an employee's trust.

But how can you do that before it’s too late?

💡 Practionable Takeaway

The next time you’re about to leverage AI for important work, ask yourself:

➡️ If AI got this wrong, would I catch it, or would I instinctively trust the output?

➡️ If AI got this wrong, and I didn’t catch it, what would it cost me?

➡️ Is there a piece of this I should do myself, to keep my judgment sharp?

Don’t throw everything to AI blindly. Use these questions to ensure the lines between your tasks and your judgment aren’t blurred.

That’s self-leadership in action.

🎥 Want to Go Deeper? 

Cognitive scientist Vivienne Ming just published a book on exactly this problem, and she's got a name for what happened with James. She calls it the difference between "automators," people who take AI's output and run with it, and "cyborgs," people who push back, question it, and keep their thinking engaged.

The cyborgs in her research consistently outperformed both AI alone and people working without it. The automators did not. James, in that moment, was an automator. He’s since become a cyborg.

If you want the fuller picture, including a practical prompting habit she calls the "Nemesis Prompt" for keeping your own judgment front and center, UC San Diego's interview with her is worth a read >> What Skills Do Humans Need to Become Robot Proof in the Age of AI?

For me, it’s yet another reminder that self-leadership is what will keep us human in an age of AI.

To Your Human Success,

Laura 💜🧡