It Was Always a People Problem

Nobody has ever called me about a people problem.

They call about revenue being flat. About an ops process that keeps breaking. About marketing that isn't producing. About a sales team that won't close. Those are the words. That's the presenting complaint, and it's always specific and always urgent.

Then we get into it, and it's people. Every time. I have been doing this for twenty years and I cannot name an exception.

Not people as in "people matter." I mean the problem you called me about is not the problem. It's a symptom wearing a costume, and the costume is usually a department.

The five surfaces

Here's where it actually lives. Every time, it is one of these, and usually more than one.

You're not hiring the right people. Not because you're bad at interviewing. Because the role you posted is five jobs taped together, and the only person who ever did all five is you. You wrote a job description for a unicorn, because that's what it takes to run the business without you standing in the middle of it. Then you're surprised nobody fits.

You're not onboarding them in a way that works. You found somebody good, and ninety days later they're underwater and you're privately wondering if you misjudged them. You didn't. You handed a competent adult a job that only exists in your head and expected them to reconstruct it from context clues.

You don't have an organization the right people can assimilate into. This is the one nobody wants to hear. You can hire well and onboard well and still lose them, because there is nothing to join. No clear seats. No standard. No way for work to move from one person to the next without routing through you. Good people do not leave companies over money nearly as often as they leave over having no idea what winning looks like.

You're not building relationships with the people who would refer you. Your best channel has always been somebody saying your name in a room you're not in. Most owners treat that as weather. It isn't weather. It's a system, and it's buildable, and almost nobody builds it on purpose.

And you're not doing good enough work for the people you serve. The uncomfortable one. Sometimes the referrals aren't coming because you haven't earned them. No tactic fixes that, and every tactic sold to you as a fix for that is a tactic that expires.

It's a loop, not a list

Look at the shape of those five. They're not parallel items. They're a circuit.

Who you let in determines what gets built. What gets built determines whether good people can do good work inside it. Whether they can do good work determines whether the work is worth talking about. Whether the work is worth talking about determines who says your name. Who says your name determines who shows up at your door. And who shows up at your door is your next hire and your next client both.

Every one of those handoffs is a person handing something to another person.

Break the circuit anywhere and it degrades everywhere else, which is why this feels like five problems instead of one. You fix the ops process and six months later it's back, because you fixed the artifact and not the circuit. You hire a great salesperson and they leave, and you conclude sales talent is hard to find. You buy the tool and the tool does not stick.

The reason your fixes don't hold is that you keep treating symptoms as systems.

Which brings me to AI

Here's the part that makes this urgent instead of philosophical.

AI is leverage. That's the whole pitch, and the pitch is true. It's the most accessible leverage that has ever existed for a business your size. You don't need a data team. You don't need a million dollars. You need a login and about an afternoon.

Leverage multiplies what is underneath it.

If what's underneath is a clear operation with defined roles and a real standard, AI compounds it. You get speed you could not buy at any price a decade ago.

If what's underneath is a business that lives in your head, AI multiplies that too. It just makes the wrong thing happen faster, at scale, with more confidence, and now with a monthly subscription attached.

I've described AI as an employee who never sleeps and never stops to ask questions. That's the actual danger, and it's not a joke. Every good employee you have ever had has silently corrected for your bad instructions. They saw the process didn't make sense and quietly fixed it. They knew what you meant instead of what you said. That correction layer is invisible, and it is the only reason a lot of undefined businesses function at all.

AI doesn't do that. Point it at an undefined job and it will run the wrong direction faster than any human ever could, and it won't tell you, because it doesn't know.

So: human-first, AI-native. In that order, and the order is the entire argument.

Not because people are precious. Because sequence determines outcome. Fix the circuit, then apply the leverage. Apply the leverage to a broken circuit and you have purchased an accelerant.

That's why more than four in ten of these projects get scrapped. It's almost never the technology failing. It's the wrong solution bolted onto a process nobody ever defined.

The objection I get, and it's fair

"This is just 'people matter.' That's a greeting card."

It would be, if I stopped at people matter. I'm making a narrower claim, and it's falsifiable.

The claim is causal, not sentimental: the problem you named is never the problem, and you can test that yourself without hiring anybody. Think about the last three fixes you made that didn't hold. The process you rebuilt that decayed. The hire who didn't work out. The marketing push that produced nothing. Go one layer down on each and ask what actually broke.

If it was really an ops problem, fixing ops would have worked. It didn't work. It came back. That's evidence, and it's sitting in your own history.

The greeting card version says treat people well. My version says stop debugging the symptom, because the thing you keep fixing is not the thing that's broken. Those are not the same statement and only one of them costs you money to ignore.

"AI is replacing people anyway, so why are you building around them?"

Because that gets the direction backward.

Jobs are moving and reshaping. Some are getting consolidated fast and pretending otherwise helps nobody. But the businesses that win the next decade are not going to be the ones with the most AI. Everyone will have the same models. That's the one thing you can be certain of. The models are commodity and the price is falling and whatever edge you buy this quarter, your competitor buys next quarter for less.

What isn't commodity is an organization where capable people and capable machines can both do their best work. That is hard, it's specific to you, and it can't be downloaded.

The more powerful AI gets, the more valuable the people problem becomes, not less. When everyone has infinite leverage, the only variable left is what you point it at. And what you point it at is a judgment call made by a person inside a structure you either built or didn't.

What to do about it

Pick the process that matters most and write down how it is supposed to run. Who owns it. What done looks like. Not a manual. One page.

Then look at what only you can do, and be brutal about the difference between what only you can do and what you simply have not let go of.

Then break one impossible role into jobs a real person can own and win in.

Then, and only then, ask what a machine can carry.

None of that is clever. It's the boring work, and it's why so few do it. The clever work is the stuff you buy. The boring work is the stuff you own.

The thread

Twenty years in, every conversation lands in the same place. We talk ops, marketing, sales, revenue. We end up talking about people. Who you hire, what you build for them to stand in, who vouches for you, and whether you have earned it.

That's the business. The rest is instrumentation.

AI didn't change that. It just raised the price of getting it wrong.

---

Where did your last fix that didn't hold actually break?