The Dependence Economy
The global economy has run on a single idea for about a century: capture attention, monetise it. Radio figured it out. Television scaled it. The internet blew it wide open. Google turned intent into an auction. Instagram turned idle minutes into ad inventory. The mechanics changed; the game didn't. Get in front of people, hold them there, sell something.
That game is breaking.
AI doesn't fight for attention the way a website or an app does. It sidesteps the whole thing. Ask ChatGPT a question and you don't get ten blue links. You get an answer. Soon you won't even get an answer — you'll get a completed task. The gap between wanting something and having it done is shrinking fast, and the old funnel of search-click-compare-decide is collapsing into something much shorter: ask, get, move on.
That changes who holds power. If people stop navigating the landscape themselves, there's no point competing to be the most visible thing in it. The value shifts to whoever becomes the navigation system people rely on.
What fills that role isn't loyalty or engagement in the way marketers usually mean it. It's dependence. That sounds ominous, but in practice it's mundane. You stop opening four tabs to cross-reference things. You stop copying data between tools. You stop doing the tedious translation work between what you want and what the software needs. The AI handles it. You just talk.
Convenience at first. Then habit. Then you can't quite remember how you did it before.
The stickiness is different from anything software has produced before. A spreadsheet stores your data. An AI learns your patterns — how you phrase things, what you care about, what you skip. After six months of use, your AI has context about you that would take weeks to rebuild somewhere else. Leaving isn't like switching from Slack to Teams. It's more like losing a colleague who knew how everything worked.
In the attention economy, you paid with time. In the AI economy, you pay with dependence.
The companies building these systems know this. OpenAI, Google, Anthropic, Apple — they're not racing to be the most useful tool. They're racing to become the default layer between you and your decisions. That's a much bigger prize. A tool gets used when needed. A default layer gets used for everything.
And once you're the default, you shape what happens next. Which emails get answered first. Which suppliers get recommended. Which options even surface. Individually, these are tiny choices. Over months and years, they steer businesses and careers in directions most people won't notice until they look back.
This is what makes the shift from attention to dependence so consequential. Attention is cheap and fleeting — you can grab it with a headline and lose it in seconds.
Dependence compounds.
Once someone has rewired their workflow around an AI system, the cost of leaving isn't a subscription fee. It's weeks of lost context and muscle memory.
None of this is theoretical. It's already the default behaviour for millions of knowledge workers. People ask Claude or ChatGPT before they Google. They delegate drafts, analysis, scheduling. Each time they do, the old model loses a little more ground. There's no single moment where it tips over. It just gradually becomes obvious that the economics have moved.
For anyone running a business, the strategic question isn't "how do we get attention?" anymore. It's: where is dependence forming for our customers, and who controls it? If your buyer's primary interface is an AI, your brand isn't competing in an open market. It's competing to be selected by a system your buyer trusts more than they trust you. And if that system becomes how they think and decide, it doesn't just influence demand. It gates it.
That's not a reason to panic or resist. It's a reason to pay close attention. Dependence, once formed, is very hard to undo. The companies that own this layer won't just participate in what comes next.
They'll set the terms of the economy that follows.