The old 80/20 rule was a rule about clients: roughly 80% of your revenue comes from 20% of your customers. Vilfredo Pareto spotted the pattern in 1896, and for over a century it told founders one thing — find your best relationships and protect them.

Dermer's Rule is a different rule for a different era. It is not about which clients matter. It is about who should do each piece of work. The rule is simple: for every function in your business, roughly 80% of the work is repeatable and should be done by AI — and the remaining 20% should be done by a human, because that 20% is exactly where humans are more adept than any machine.

The old rule asked which clients matter most. Dermer's Rule asks a sharper question: within each function, which work belongs to the machine — and which belongs to the human?

The clarity is in that second half. Dermer's Rule doesn't say "use more AI." It says: the 20% you keep is not leftover work. It is the human-adept work — judgment, taste, trust, hard calls, and the reading of a room — that AI cannot do well and, for the foreseeable future, will not. Draw that line correctly in every function, and you've built an AI-native company without losing the thing that made it worth building.

Here's what makes the new version clear where the old framing was fuzzy: it doesn't ask you to guess a percentage of "how much AI." It gives you a repeatable test you run function by function — is this task repeatable, or does it need human judgment? The repeatable 80% goes to AI. The judgment-heavy 20% stays with people. Every function, same test.

The pattern is the whole point. The more repeatable and rules-based a function's output, the higher AI's share climbs. Marketing and support lead because so much of the work is drafting and first-response. Finance sits lowest — around 55% — not because the math is hard, but because the cost of an error is high and human oversight earns its keep. That's your sequencing map.

Notice what these have in common: trust, judgment, taste, ethics, reading a room. None of them are repeatable. Every one of them is contextual, relational, and consequential — the exact qualities that make the 20% human-adept. That is why Dermer's Rule doesn't shrink the human role. It concentrates it on the work that was always the point.

The 80% AI handles is the work you never wanted to do. The 20% you keep is the work only you can do well.

The infrastructure is here, the tools are cheap, and automation reclaims as much as 20 hours a week per person. The founders who redraw the 80/20 line in each function now will run at a cost base — and a speed — competitors simply can't match.

How to apply Dermer's Rule this quarter

Pick one function — marketing is usually the fastest win — and sort every recurring task into two buckets: the repeatable 80% and the human-adept 20%. Hand the 80% to AI with a human review step, and protect the 20% fiercely. Then move to the next function: support, then operations, then sales, then finance last. You're not shrinking your team; you're pointing their hours at the work where they're genuinely more adept. Do that across all five functions and you've rebuilt the company around the ratio that defines the next decade.