• AI for Leaders
  • Posts
  • Two Things that AI Leaders Should Keep in Mind (And One Must-Read Paper)

Two Things that AI Leaders Should Keep in Mind (And One Must-Read Paper)

No playbook. Start small.

Welcome to the first edition of the “AI for Leaders” newsletter!

Let’s start by highlighting two things that leaders should keep in mind as they think about how they can use AI in their organizations.

Thing #1: There’s No Playbook (Yet!)

I’ve spent quite a bit of time reading and learning about AI, especially its use in businesses.

And it’s pretty clear. No one has all the answers yet.

Companies and their leaders are still figuring things out as they move along.

So for now, there’s no set playbook.

A word I keep hearing from different thought leaders is “experiment, experiment, experiment.”

In other words, experiment with AI tools and try out different things to see what works and what doesn’t.

So if you’re feeling a bit of FOMO, don’t worry.

It’s still early, and everyone is exploring what’s possible.

Thing #2: Start Small

Jumping into AI can feel like there’s a lot coming at you.

There are tons of technologies coming out every day, and it’s easy to get overwhelmed and confused by all of it.

Take a step back and start small.

Here’s one idea:

  1. Go to chatgpt.openai.com.

  2. Sign up (it’s free)

  3. Ask for help on something small, like rewording an email, or drafting a meeting agenda

Don’t overcomplicate it.

If you’re still feeling uneasy, the free book I shared with you in the welcome email (ChatGPT for Better Business Communication) gives you everything you need to get started.

If you didn’t download it yet, click the button below to subscribe and grab the book 👇️ 

One Interesting Article

In every edition, I’ll share an article that I think could be helpful for you as a leader.

The following working paper, written by a group of social scientists and thought leaders from the Harvard Business School, Boston Consulting Group, Wharton School, and MIT Sloan has been an eye-opener about how AI will help businesses.

You can download the full paper by clicking here 

It is 58 pages long (i.e., not a short read), so here are some key takeaways from it:


  • The study examined “performance implications of AI on realistic complex, and knowledge-intensive tasks” (they focused on 18 different tasks)

  • It included 758 consultants from Boston Consulting Group

  • The consultants were assigned three different conditions:

    • No AI Access

    • GPT-4 AI Access

    • GPT-4 AI access with a prompt engineering overview


  • Consultants who used AI were “significantly more productive

  • On average, they completed 12.2% more tasks and did so 25.1% more quickly

  • On average, they produced 40% higher quality results when compared to the control group

  • All consultants benefited from AI, but those who were below the average performance threshold benefited more than those were were above it

    • Consultants who were below: increased by 43%

    • Consultants who were above: increased by 17%

Those results are super-impressive, but there is one important caveat to consider here.

Important Caveat regarding the “Jagged Technological Frontier”

A word of caution about those results is that they were for tasks within what the authors called the “jagged technological frontier.”

What is the “Jagged Technological Frontier?”

Ethan Mollick (a Professor at Wharton and one of the authors of the study), explains what the authors meant by that here.

In simple terms, think  of tasks that AI can easily do, and tasks that AI cannot easily do.

The frontier is a line that separates those two.

If you are within the frontier, then AI works great as a tool (produces accurate results).

If you’re outside of the frontier, then AI is not-so-great as a tool (produces inaccurate results).

The problem is that this frontier is “jagged” and invisible (i.e., some tasks that you think AI would be good at turns out to be untrue, and vice versa).

Here’s a screenshot from the paper that explains this visually.

The frontier is also growing every day, so it’s hard to nail down where that border is.

The important thing to note is that the positive results shared above were only for tasks within the frontier.

For tasks outside of the frontier, the results were flipped: “…consultants using AI were 19 percentage points less likely to produce correct solutions compared to those without AI.”

So basically, with AI tasks that AI is not-so-great at, it was better for consultants not to use it.

This is an important takeaway for leaders.

Although AI seems very promising, we have to be careful about its limitations.