What I learned about delegating to people
Most people are terrible at delegating. I don't claim to have it all figured out, but I am a lot better than I used to be.
Years ago I figured out that when I hire good people, especially people who specialize in an area I don't, the best thing I can do is not overmanage them. I give them direction on the results I am looking for, then I stay out of their way and let them do what they do.
The more step by step direction you give someone like that, the more you limit them to your own skill set. The whole reason you hired them was that they know more than you about that area.
Delegation shows up in the numbers
This is not just my opinion. In 2014, Gallup studied 143 CEOs from the Inc. 500 list of fast growing private companies. The CEOs with high Delegator talent posted an average three year growth rate of 1,751 percent, 112 percentage points higher than those with limited or low Delegator talent. They also generated 33 percent more revenue in 2013: $8 million versus $6 million.
Gallup called delegating a huge management challenge for entrepreneurs. That matches what I see with fractionals and consultants. Letting go is hard, and it costs you when you don't.
Now I am learning the same thing about AI
A lot of people are hung up on giving AI long, very specific directions about how to do things. That probably made sense six months or a year ago, depending on the model. Now part of the trick to getting the most out of AI seems to be telling it less about the how.
OpenAI says something similar in its own guidance for its reasoning models. It says those models perform best with straightforward prompts, that telling them to "think step by step" may not help and can sometimes hurt, and that you should be very specific about your end goal. Clear on the result, light on the steps. That is good delegation.
Ask for the path, not just the report
Here is how that looks in client work. Instead of only telling the AI to write up how we will do X, Y and Z to reach a goal, I give it everything it has about the client and the goal, then ask: how would you recommend we get from where we are to where we are trying to be?
That doesn't mean I take the advice and run with it. I still bring my own plan and my own judgment. But when I put my plan next to its recommendation, I get something better than either one alone.
So my advice to other fractionals is simple. Treat AI the way you would treat a strong specialist you just hired. Tell it what success looks like, give it the context, and see what it comes back with before you start telling it how.
