McKinsey published an article with a title that sounds like a promise: "Why AI will elevate instead of replace commercial teams." It's an interview by partner Alexander Dierks with Monique Buch, chief commercial officer at Covestro, and the thesis is simple. AI removes the routine work from sales teams, delivers better data on each customer, and what's left, the human part, the relationship-building, becomes the true differentiator. A salesperson who no longer fills out forms has more time to think about the person on the other side of the table.

It's a reasonable argument and it isn't false. AI is a tool that does exactly what its name suggests: it automates. What gets automated and what happens with the freed-up time depends on human decisions. That's the atomic definition of this entire piece: AI is a tool whose effect on employment depends on who controls the budget, because the same automation that frees up hours for relationships also frees up payroll for cuts.

In November 2025, while that article was circulating, McKinsey cut around two hundred positions in its technology division. The firm relied, according to reports from that time, on AI automation, and there was an internal review of which support tasks that technology could fully take over. Its internal tool, called Lilli, already does work that used to fall to junior analysts: building presentations, summarizing information, structuring deliverables. Same month. Same company. Two stories that don't contradict each other but also never appear together in any press release.

I've spent the past few years building systems to simplify life for sales teams, to the point where the ultimate goal isn't for people to update the CRM but simply to confirm that what the system already entered is correct. I agree with Buch: removing administrative work does elevate the role of the person selling. A salesperson who doesn't lose the afternoon entering data has more mental space to listen to a customer. That's not empty corporate propaganda; it's something I've seen work. Which is exactly why I can point out, with more precision, what's missing from McKinsey's framing: elevating the role and cutting the headcount aren't alternatives. They're the same automation, measured from two different budgets.

The Generosity in the Doorway talks about a pattern that repeats itself in nearly every proposal for social change: the design answers what to do, but stays silent on who decides and who gets to say no. That book examined universal basic income and resource-based economies, but the process is identical here. McKinsey answers, with precision, the question of what AI can do for a sales team. It does not answer, nor does it need to in that particular article, who decides whether the freed-up hours turn into more quality calls or into fewer people on payroll.

Why doesn't McKinsey mention its own layoffs in an article about how AI elevates work? Not because it's lying. Both things, the elevation of sales at Covestro and the technology cuts at the firm itself, can be true at the same time, and they probably are. The problem isn't falsehood, it's selection. A consultancy that sells AI transformations has a natural incentive to tell the half of the story that sounds like progress and leave the other half in an HR report that nobody turns into a headline article. It's the same process I documented with Jensen Huang redefining what counts as artificial general intelligence: whoever controls the vocabulary also controls which part of reality makes it into the frame and which part gets left out.

What does this mean for someone who doesn't work at McKinsey or Covestro, someone who simply uses AI to do their job? It means the right question is never "Will AI elevate me or replace me?" That question has no answer because it depends on a variable the technology doesn't control: who holds the budget and what they decide to do with the time automation frees up. If the sales manager decides to invest those hours in deeper customer relationships, AI elevates. If the CFO decides those hours no longer need to be paid for, AI replaces. The tool is identical in both cases.

What the news of McKinsey's layoffs adds to the book's thesis is evidence that not even the companies that best understand automation, because they literally sell it as a service, manage to resolve that tension publicly, or perhaps don't even try. The firm that explains to Covestro how to elevate its sales team is the same firm that, behind closed doors, uses the exact same logic to cut heads. There's no necessary hypocrisy in that. There are simply two different budgets picking up the same tool and arriving at opposite decisions, each rational from its own incentive.

What does challenge the thesis, or at least complicate it, is that automating routine tasks is genuinely good for the person who used to do them. Nobody misses filling out forms. Buch's argument isn't empty propaganda: removing mechanical work and providing better data really can make a job more interesting. The question was never whether automation improves the task. The question is what happens to the person once the task they used to do no longer exists, and someone else, with more decision-making power, determines whether that person stays in the conversation or leaves the payroll.

I don't have a clean answer for how to resolve that asymmetry, and I'd rather say so plainly than pretend to a solution I don't have. What I can say is that this pattern repeats with enough consistency that it stops being an isolated case: Jensen Huang defines what counts as artificial general intelligence in a way that benefits Nvidia, Anthropic and Mustafa Suleyman debate whether Claude is conscious without it ever being clear who benefits from each answer, and now McKinsey honestly documents half of its own automation logic. Stones don't lie, but corporate reports sometimes tell only half the story, not because they lie but because the other half lives in an HR press release that nobody reads with the same attention.

Which part of that story do we choose to ignore?

Sources:

1. McKinsey & Company, "Why AI will elevate instead of replace commercial teams" (interview by Alexander Dierks with Monique Buch, Covestro)

2. Reports on technology staff cuts at McKinsey, November 2025, and the role of the internal tool Lilli

3. Yves Laurent, "The Generosity in the Doorway" (Chapter 14 — The Question Beneath the Question), Amazon ASIN B0H6RT5Y32