It's 4pm and something broke in last night's release. The rollback is staged. It's sitting there, ready, waiting for somebody to say go.
Your tech lead knows exactly what broke and exactly what a rollback costs in engineering hours. What your tech lead might not know is that the same release carried the one feature you promised the board would be live by Thursday, and that the thing that broke is a secondary feature no customer has touched yet.
Roll back and you're technically correct and you've just pulled your own board demo. Don't roll back and you're knowingly shipping with a defect for two days.
Sometimes this is easy. The app is down, roll it back, nobody needs a strategy conversation about it. But the calls that actually cost you money are never the easy ones. The hard version needs somebody holding two things at the same time: enough technical grip to know what the rollback really does, and enough business context to know what Thursday is worth.
That person may not be in the channel. Nobody notices, because the decision still gets made. It just gets made by whoever happened to be closest to the keyboard, using half the information.
That's an Operator Vacuum. Not a vacant role on an org chart, just nobody holding both halves of a decision at the moment the decision has to be made.
A Very Particular Set of Skills
(Yes, I'm doing the Liam Neeson voice. Let me have this one.)
Here's the thing though. The reason the Operator Seat is hard to staff is that it asks for a combination most orgs never assemble in one person. You need someone who can talk to a customer and actually hear what they need, turn that into work with a real priority attached, carry enough standing with engineering that the work gets done rather than politely queued, and then go back to the customer and close the loop. In most companies that's two or three different people. It's one decision.
This isn't a thing I made up to describe my own résumé. John Levisay, a former CEO who now works with venture and PE-backed companies on talent strategy, described the same problem in April for Hunt Scanlon Media, and he was talking about marketing hires rather than delivery. He said the role usually needs someone who can "write copy on Tuesday, build the dashboard on Wednesday, fix funnel leakage on Thursday, and still recruit on Friday." He calls that person an operator (which I swear I did not pay him to do).
His piece is called Companies Don't Have a Talent Problem, They Have a Judgment Problem. He describes his own hard-won lessons as "part confession, part scar tissue, part operating manual." I've been calling my own version twenty years of scar tissue for a while now, so reading a stranger reach for the same word was a strange afternoon.
Judgment is earned, not taught. That's the part you can't shortcut, and it's the part every plan to fix this cheaply skips right over.
I'm not smug about this. I didn't have that judgment at thirty-two, and I'm not sure I had much confidence either. What I had was a lot of runway left and a high tolerance for being wrong in public, which turns out to be how you get the other two.
The Sorting Happened Before You Met Them
The normal route to a leadership seat requires shedding the tactical layer as you climb. You get promoted precisely for handing the hands-on work to somebody else.
Your company probably doesn't have much of a ladder. You're thirty people, the titles are approximate, everybody does three jobs. But every candidate who walks through your door came off somebody else's ladder, and that ladder did its sorting years before you met them. Levisay describes companies hiring an impressive name who "has not personally been in the weeds in ten years," then acting surprised when nothing ships. Ten years is roughly how long the sorting takes.
So the pool you're hiring from has already been filtered, and not by you. Not by anybody's incompetence either. By the career track working exactly as designed.
I came up doing delivery work, and then I left to go do something else entirely for a few years. Van-lifed around Europe. Ran an Amazon business. Stayed a long way from anyone's org chart. When I came back to tech it was as a freelancer, which meant there was never anybody in a position to promote me out of the hands-on half. I didn't hold onto both halves through discipline. I stepped off the track that takes one of them away.
So You Buy Something Instead
Which brings us to the ads that have been popping up in my feed lately.
Let me be honest about where I'm standing before I say anything else. I would probably use these tools. I already run a homegrown version of one, just scheduled tasks and a pile of context, no venture funding and no interface to speak of. I think I spend more time with Claude than with my own husband. When I got this wrong early on, the problem was me: I was not being a good boss to my AI employees. So this isn't a complaint about the software. The software is good.
I'm going to pick on one company below, and only because they say it most plainly. There are a dozen of these now and the pitch barely changes between them.
Viktor raised a $75 million Series A this spring and is advertising to founders with the line "Not a tool. A hire." Underneath that: hire Viktor and run your company like it's twice the size. There's a whole section headed Employee, not software, and the first thing under it is Takes ownership.
Scroll down a little further, to the part where they explain how you stay in control, and the same page says: He proposes, you decide. And: every sensitive action waits on your approval.
Both of those claims live on the same page. The second one is the true one, and it's to their credit that they put it there, because gating the irreversible stuff is the responsible way to build this. But read it as a founder for a second. The thing they've carefully not automated is the deciding.
Their own demo makes the point better than I can. In it, Viktor catches a checkout outage, traces it to last night's deploy, opens the incident, pages on-call, stages the rollback. Then it says the rollback is staged and waiting on your go. A human named Sophie has to say the word. They wrote that demo. It's their best foot forward, and the only judgment call in the whole thing belongs to a person.
So look at who that person has to be. Somebody who knows what the deploy actually broke and what it's actually worth, holding both, in the next few minutes, under pressure, and carrying enough standing that if they need to say "ship it anyway" nobody walks it back in a DM an hour later. That's the Operator Seat. Viktor doesn't fill it. Viktor is very good at telling you it's empty.
Noticing is not deciding. Every one of these products is excellent at the first one and hands you the second.
I've now watched four different companies build variations of the same thing. One of the founders told me on a call, without any prompting, that his product doesn't make judgment calls and the decision stays with the operator. He meant it as a feature. He's right that it's a feature. It's also a complete description of where the product stops.
Tools are not the solution. They enable a solution you've already defined. If you haven't defined who decides, a very good tool will escalate very efficiently to a person who can't answer.
That's the Tool Trap, and this version is harder to spot than the one I wrote about back in June, because nothing about it looks like avoidance. It looks like diligence. You bought good software, you connected it properly, the escalations are cleaner than they have ever been. The gap they're escalating into hasn't moved an inch.
Maybe Somebody Is Already In the Building
The answer to "can't someone here just do this?" is sometimes yes.
If you have that person, they're probably a level or two below where you've been looking. The ladder hasn't stripped them yet, which is exactly what makes them useful and exactly why they don't have the title. And they're likely the person everyone already routes around, because holding the seat without any authority attached to it is a job nobody can actually do. I wrote about that last week. Naming an owner and giving them the standing to decide are two different acts, and most orgs only perform the first one.
So before you conclude you need to hire anyone, go look. Not at who owns the system. At who in that 4pm channel has the system, the calendar, the strategy, and the customer in one head. If the answer is nobody, your tooling did its job perfectly and your org still can't answer the question.
And a warning about what happens if you do fill it, inside or outside. People get tense. That tension is not evidence you chose wrong. It's the sound of the seat finally being occupied, and if you back off the first time it gets uncomfortable, you'll be right back in this channel in six months with a shinier tool and the same problem.
Which is a different conversation, and I suspect the harder one.