I will always remember the day I met the CFO of the company I worked for in my first post-MBA role. I’d been given the task of putting together some type of report for him, and I was shocked when I was told to deliver it to him directly. Upon entering his office on the top floor, I was naturally nervous. The report I’d been asked to put together hadn’t made much sense, so what I was delivering was a quite different version that I thought worked better. When I handed it to the CFO, he sat down with it and asked me several questions on it. With each question I became increasingly nervous. Then he said that he wanted to share the first piece of real career advice he had received. He told me

“Don’t always give your boss what they want. Give them what they need.”

Clay Creasey

He gestured at the report I’d given him and said that I’d done the right thing in making it more useful. I was relieved, and I was emboldened to repeat that success.

That pithy line encapsulates the biggest advantage of humans over robots or programs. As anyone who’s written even a few lines of code can attest, the best and worst thing about computers is that they do exactly what they are told. Humans have the gift of vetting every instruction in the light of more generalized intelligence. Programs do not.

For that reason, the hashtag #shutdowntheAI started trending on Twitter in the data science community a few years ago. Here’s one example of how things can work out when instructions are followed to the letter:

In my game, the town guard wants food and to punish criminals, so he’d steal food from the tavern then kill himself. I had to #shutdowntheAI

— Stephen Ware (@sgware) August 1, 2017

Another example shared elsewhere was that of someone who had constructed a stick figure to represent a human, and who had instructed the AI to obey the laws of physics while moving to a set finish line. The idea was to see if the AI could learn how to walk. Since instructions needed to be explicit, the goal was specified as having the stick figure’s head cross the finish line. The AI found the simplest solution, which was to have the stick figure sever its own head and throw it across the finish line. In a second pass, the constraint of having all body parts connected when the head crosses the finish line was added. In that one, the AI had the stick figure dismember itself and reattach each body part end-to-end until it was simply a head on a tower of body parts, then fell forward so the head crossed the finish line.

These are obviously extreme examples of doing what one is told rather than what’s actually needed, and no actual person would go to this extreme (or, if they did, they would soon no longer be a problem). Still, they highlight that the biggest advantage of having real people work for you is that you can trust them to use their brains in a wider variety of ways than what you specifically demanded. Employees who are adept at exercising this discretion are invaluable.

If you are inclined to follow this advice in your career, make sure you take care when exercising this discretion. It can take a certain amount of arrogance to assume you know better what needs to be done than the one who told you what to do. Any time you substitute what you think your boss needs for what they want, you’re substituting your judgment for theirs. Tread lightly, and make damned sure you’re right.

The higher the stakes, the more important it is to do what is needed, and also to ensure you’re correct in your judgment. Whenever possible, communicate these decisions before acting on them. Sometimes, you may communicate your decision and lose the argument, being told to do what you were originally told and not follow your judgment instead. In those cases, proceeding forward with what you believe is needed anyway is incredibly hazardous, even if you turn out to be right. Know what risk you are taking.

This advice does not just apply to early leaders. It applies doubly to senior leaders, and it even applies to executive leaders, though usually on the reverse side. For years I have shared this advice with my direct reports and with those under them. It is dangerous and important advice to give to people under you. It is anathema to command and control leadership styles. It opens you up to what can later be viewed as insubordination. It widens the gap between what you expected or asked for and what you receive. And it has the potential to transform your career.

In 2018 I had the pleasure to see General Stanley McChrystal, author of Team of Teams, present. In his presentation he phrased this same advice as

“If, when you get on the ground, you find that the orders we gave you don’t make sense, exercise the orders we should have given you.”

General Stanley McChrystal

General McChrystal understands that in business, war, and life, ephemeral knowledge is a real thing. As a leader, it is vital to know and understand your area of the business. Still, it will always be impossible to understand every area of the business in the detail that someone immersed in that area at that time can. Giving your employees your blessing to exercise their judgment when necessary can save you from threats you weren’t even aware existed.

For this advice to be effective in the hands of your employees, you must give immediate, specific, direct feedback (whether positive or negative) every time you see your people doing this. Encourage good judgment, discourage rash decisions and tunnel vision, and always strive for the most effective communication possible. Note, again, that in order for this advice to ever have value, employees must always have an understanding of what vision they are striving toward. If they don’t know the goal or understand why they’re working toward it, they can never contribute beyond doing what they were told.

Sometimes, there may be something so important or specific that you cannot afford any substitutes for your judgment. You can avoid being undercut or blindsided in these scenarios simply by telling your people that this task is to be done exactly as stated. Give the reasons if you can. If you can’t, then just tell them to trust you that there are good reasons for following your directions exactly in this case, and that you’ll make those reasons clear to them when you’re able. If you establish consistency in giving reasons where you can and in responding well to your people making these judgment calls, they will in turn trust your judgment on this.

In addition to General McChrystal, the leader of a Fortune 5 company also had a good way to phrase this advice:

“It doesn’t make sense to hire smart people and tell them what to do; we hire smart people so they can tell us what to do.”

Steve Jobs

If giving the above advice to your employees scares you, ask yourself why that is. Do you really trust your judgment over theirs, even in cases where they may be closer to what’s going on than you are? If so, why aren’t you hiring smarter people? An entrepreneurship professor I had in undergrad once quipped that “A level people hire A level people; B level people hire C level people.” Did you hire A level people or C level people? What does that say about you? If you hired C level people, the first step is admitting you have a problem. If you hired A level people, capitalize on what you hired them for- their ability to think.