Sample frameDay 148·AI Fluency · Redesign

Morning. Your process was built around a scarcity that has quietly stopped existing.

AI Fluency.

Redesign, Don't Bolt On

5 min read·Apply by lunch

The question

How many steps in the process you've just made faster with AI exist only because producing a first version used to cost a person a day? Count them before you answer, because the number is usually most of them, and every step you kept is a decision you've never actually made.

The idea

Michael Hammer, a former professor of computer science at MIT, opened his 1990 Harvard Business Review article Reengineering Work: Don't Automate, Obliterate with a complaint about technology spending: companies were investing heavily in computers and weren't getting much back, because they were mechanising the processes they'd already built. His example was Ford's accounts payable operation in North America, which employed more than five hundred people. The plan had been to automate it and remove a fifth of them. Then somebody looked at Mazda, a far smaller company, whose equivalent function ran on five. Ford stopped automating, redesigned the process itself around the receipt of goods rather than the matching of paperwork, and took out roughly three quarters of the roles. The trap is the automation that reports well: the same work, measurably faster, a clean number for the board — and the redesign that would have removed the work entirely never reaches the agenda. Hammer's argument was that the shape of a process encodes the constraints of the era that produced it. Automate it and you buy the old shape, faster and at greater expense. Redraw it and you get a different shape entirely.

The framework

01
Every step in there was rationing something.
Walk the process step by step and answer one question for each: why does this exist? A striking number resolve to the same answer, which is that producing a first version used to be slow and expensive, so the organisation rationed it. The briefing meeting exists to get the draft right first time. The template exists so that nobody starts from nothing. The queue exists in front of the one person who writes well. The two-week lead time exists because that's how long the queue takes. None of those steps was stupid. They were sensible answers to a constraint that has since moved.
02
Redraw it assuming drafts are cheap and review is the bottleneck.
Start again on a blank page with one assumption changed: a first version costs minutes, and the scarce resource is a qualified person's attention. The shape usually inverts. Instead of one carefully briefed draft arriving in a fortnight, three sharply different options arrive on Tuesday and the meeting becomes a choice rather than a briefing. Instead of protecting the writer, you protect the reviewer: fewer things reaching them, arriving in a form that makes the decision quick. Do this on paper before you open any tool, because a redesign performed inside a product's interface will only ever end up the shape of that product.
03
Move the drafting to whoever needs the output. Keep the judgement where it is.
Hammer's sharpest principle was to have the people who use the output of a process perform the process, which was impossible while the work demanded a specialist and frequently isn't any more. The account manager who used to brief the analyst can now produce the first cut and bring the specialist a question instead of a request; two handovers and most of a week go with them. Be honest about the boundary. This works where the specialist's contribution was polish or speed. It fails where the contribution was judgement, and pretending otherwise is how a redesign turns into an incident.
04
A fortnight on live work, with the review date already booked.
A redesign that stays in your notebook is an opinion. Write the old process and the new one side by side, count the steps and the handovers in each, name what has been removed and who owns whatever replaced it, then run it on live work for a fortnight before anyone is allowed to argue. Set the review date in the same sitting, the way Day 44 has you schedule learning rather than wait for it, because a redesigned process fails in ways the old one never did and you want that surfacing in a booked hour, not in front of a client.

Members only · 4 principles + template + AI mentor

364 more frameworks are waiting.

Redesign, Don't Bolt On is Day 1 of 365. One framework every morning for a year — across strategy, sales, negotiation, leadership, and more.

$1 /day

Billed $365/yr · cancel any time

365 frameworks

12 topics · templates · AI mentor

Reserve your place — $365/yr →

One payment a year · the whole library · no auto-renewal.

More AI Fluency frameworks

All 20 ai fluency frameworks →