The development of full artificial intelligence could spell the end of the human race.
AI is not going to replace humans; it's going to enhance human capabilities
My smartwatch told me to breathe. I’ve been breathing all my life, but sure, I’ll take advice from my wristwatch.
Two years ago I wrote a piece on this site arguing that AI was moving from the logic layer to the touchpoint layer. For a decade, AI and ML had been embedded inside the products we used every day, invisible inside ranking algorithms, ad targeting, recommendation feeds, photo processing. Then ChatGPT happened and suddenly AI wasn’t underneath the product. It was the product. The interface itself became the thing.
That argument has aged well.
Look at what’s happened since. Conversational AI has become the default interface for productivity software. Voice interfaces are good enough to use seriously for the first time in their history. Agentic systems are quietly taking on tasks that used to require navigating a UI. The whole stack we used for thirty years, click menus, fill forms, navigate pages, is being replaced by talking to a system that does the work for you.
The touchpoint shift is now obvious.
What I didn’t see clearly two years ago, and what’s clear to me now after a year of building production AI at LOST iN, is that the touchpoint shift was the easy part. The deeper shift is happening one layer below, at the operating system level of the company itself.
Here’s what I mean.
When AI was just inside the logic layer, companies stayed the same. The org chart didn’t change. The processes didn’t change. AI made existing workflows faster or smarter, but the workflows themselves stayed roughly the same.
When AI became the touchpoint, things shifted slightly. Customer service teams shrunk. Content creation teams restructured. Some workflows changed shape. But the company underneath was still recognizably the same kind of company.
What’s happening now is different.
Companies built around AI from the start, not retrofitted with it, are starting to look structurally different. Smaller teams. Different roles. Workflows that don’t have human handoffs the way old workflows did. Agents handling work that used to require entire departments. The operating system of the company itself is being rewritten.
At LOST iN we have a small team, a production AI concierge, an agentic content pipeline, and infrastructure that handles what would have been a 30-person operation two years ago. We don’t have a customer service team. We don’t have a content moderation team. We don’t have a research team in the traditional sense. Agents do most of that work. The humans are higher in the stack, doing things agents can’t.
This isn’t unique to us. The pattern is showing up everywhere I look. The companies designed around AI from day one are leaner, faster, and structured around a small number of operators who can write code and steer agents. The companies retrofitting AI onto a legacy structure are running into the same problem repeatedly. The structure fights back.
What this means for operators.
Every senior leader I talk to is trying to figure out what their company looks like in three years. The honest answer is most don’t know. The org charts they’re working from were designed for a world where humans did the work and software made them faster. The world we’re building has agents doing the work and humans steering them. Those are different shapes.
The leaders who navigate this well are going to be the ones who internalize that the touchpoint was the appetizer. The real meal is structural. New role definitions. New team shapes. New cadences. New ways to think about what a company even is when the operating layer is no longer human.
I’m building one of these right now. It’s the most interesting work I’ve done in twenty years, and it’s also the hardest. Not because the technology is hard. The technology is the easiest part. The hard part is everything around it. Hiring people who can think this way. Designing systems that compound learning across agents. Knowing when to defer to AI and when to override it. Building trust with customers when the interface doesn’t have a human on the other end most of the time.
That’s the work.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Two years ago I argued AI was becoming the new touchpoint. That call held up. The thing I’d add today, and the thing that most strategy memos still aren’t naming clearly enough, is that the touchpoint was always going to be temporary. It was the most visible part of the shift, but it wasn’t the actual shift. The actual shift is the company itself.
That’s the part I’m writing about now, because that’s the part I’m building.