Mental Model #01 · Mental Models for the AI Era

Most companies store intelligence. Almost none of it makes it flow.

‍

Walk through your own company and look at where the knowing actually lives.

‍

It sits in the heads of your best people. In reports nobody opens twice. In the one analyst who answers when asked. In a data warehouse that took eighteen months to build and gets queried by four people.

‍

Systems theory has a name for this: a stock. A quantity that piles up in one place and waits to be fetched. And fetching is what your company actually runs on.

‍

The Monday trading meeting exists to move numbers into people’s heads by hand. The board pack exists to move a month into a room, weeks after the month ended. Every “let me pull that and get back to you” is a bucket carried up the stairs. Nobody designed it this way. It is simply what intelligence logistics look like when people are the only carriers available.

‍

The cost hides in plain sight because it books no invoice. It surfaces as latency: decisions made on last month’s picture. As queues: the business waiting for the one person who owns the numbers. As altitude: your most senior people spending their sharpest hours assembling the past instead of deciding the future.

‍

When that person goes on holiday, the company gets measurably less intelligent. That sentence should be absurd. Check whether it is true of yours.

‍

Donella Meadows, the great systems thinker, drew the distinction this whole model rests on: stocks sit, flows move.

‍

Water in a tank is a stock. Water in the pipes is a flow.

‍

For a century, business intelligence has only ever been available as a stock, because moving it required a human being to pick it up and carry it. Every ritual you run, the weekly review, the reporting pack, the steering committee, is a bucket brigade built around that constraint.

‍

When installed at the core of a business (not at the edges!), AI changes the state of matter

‍

Not because it is smarter than your analyst, but because it never sleeps, never queues, and can be wired directly into where your business actually happens. Intelligence stops being a stock you visit and becomes a flow: it moves through the business on its own and arrives where a decision is being made, at the moment it is being made.

‍

Water in the pipes, not buckets on the stairs.

‍

This is why a cleverer dashboard changes so little. A dashboard, however intelligent, is still a place you must go, still a bucket you must carry back to the decision. A bigger, faster, more beautiful bucket. The shift that matters is not a smarter tool you consult. It is a utility you wire. Nobody “checks the electricity” before turning on a light; pressure is simply there at the switch. That is the standard the AI era sets for intelligence, and the companies that reach it first will be very hard to catch.

‍

The executive job changes with it. It stops being “accumulate more intelligence”, more reports, more dashboards, more analysts, and becomes “lay the pipes it moves through”: your data in one place, your standards written down, the flow arriving at the decision instead of the decision walking to the stock.

‍

Less collecting. More plumbing.

‍

One warning before the next model in this series: when intelligence becomes liquid, it does not trickle. It floods. Cheap, abundant cognition multiplies the questions, the options, and the decisions on your desk, and abundance is a different problem from scarcity.

‍

That is where the Jevons Paradox comes in, as Mental Model for the AI Era #2.

‍

This is the first of 10 mental models for running a business in the AI era, published every Wednesday.

‍

If you want to know how much of your company’s intelligence still travels by bucket, this AI Readiness Diagnostic will tell you in a few minutes.