If AI can make your deliverable in minutes, you’re running a DVD company. Netflix is here.

The DVD test

That is why our working rule is: evolve, don’t optimize.

The two-year question

Doing the same thing faster is a losing proposition

At a business event, I spoke with an executive from a large company in the HR space. They were using AI to interview candidates and to write better training manuals. It made sense on the surface. It was also extending the life of a DVD company.

The better question is: how do we capture what our best people know so that AI can do the job? When an experienced person writes down the actions and processes that win, that knowledge should train AI, not produce a nicer manual for the next hire. And the real value in that business sat in the proprietary data inside the companies it served. Whoever learns to train AI on that data first wins.

Your company is now intelligence

Digital intelligence is fast, but it has been trained largely on internet data, and much of that is mediocre, so its default output is mediocre. Biological intelligence holds deep, hard-won knowledge, but it is a slow server. The operator’s job is to decide where each goes: use people to capture what good looks like, so the digital side can learn your standard, then let the digital side run and keep itself current. This matches McKinsey’s finding that value comes from redesigning workflows rather than inserting AI into existing ones1.

Where value goes when the deliverable becomes cheap

It goes to the people who own and operate the business the deliverable was serving.

An agency knows how to generate demand, build a funnel, test an offer, write the copy and read the data. For years it sold that skill to someone else and they kept the asset it helped build: the customers, the data, the brand, the cash flow. If AI makes the skill cheaper for everyone, its hourly value falls. The value of the asset does not. A business with real customers, real data and a real position in a market is still worth something, arguably more, because the cost of running it fell.

There is a timing element. A service company that has adopted AI seriously has a technological edge today. That edge is temporary; in a few years everyone will have it. But it can be moved, from the know-how of the team into vertical businesses the team understands well, where it compounds instead of being billed once.

What this looked like at Monarchy