What Changes When the Slow Part of Your Business Stops Being Slow

A few years ago Jeff Watts, now chief executive of Fastenal, the industrial-supplies distributor, was visiting branches and watching how they quoted work. It upset him. "It was the same way I quoted business 30 years ago," he told analysts on October 14. Fastenal added AI and some of its internal technology to an existing quoting tool, and he gave one result: sourcing and quoting a job that used to take "5, 6 days" now "may take 15 minutes."

Watts told it as a cost story: "I mean that time savings alone." Many AI business cases are built the same way. Take the old process, time it, time the new one, multiply the difference by a salary, and you have a number for the board.

That number answers the smaller question. The bigger one is what a company does once the thing it used to ration stops being scarce. A quote that takes a week is something a branch has to hand out carefully. A quote that takes 15 minutes is something it can give to anyone who asks, and once that happens, the reasons the business was set up the way it was can start to fall away.

Quotes used to be rationed

If sourcing and quoting a job takes five or six days, a branch can't quote everything it is asked to, so something decides which requests get the time. The likeliest answer is the big account with the steady order, while smaller or odder requests wait their turn. A rule like that is probably not written down anywhere.

Count the hours saved and those habits stay in place. The branch does the same quotes for the same customers, only faster. Ask what you would do if a quote cost almost nothing, and harder questions show up. Should every request get an answer the same day? Is the small customer you used to turn away worth having now? Does the salesperson's job move from chasing quotes to choosing which answers to push? The point is that Watts described the tool by how long a quote takes, while the bigger decision may be about who gets one.

Watts gave one more detail. Branch managers who once asked for "2 or 3 heads" after landing an account now say: "We don't need the 2 or 3 heads. We need this tool." Even here, the request is framed as people saved; the open question is what the branch does with the tool once it has it.

Streaming did this to television

Sangeet Paul Choudary, who writes about platforms and strategy and is the author of Reshuffle, made the same argument about television in an August essay. "Traditional TV programming was built around a scarce asset - prime-time slots," he wrote. A network had a few slots and far more shows than it could put on air, so the programmer's job was choosing which shows deserved one.

Choudary argues that streaming's bigger change lay elsewhere: "the real scarcity it removed was the scarcity of the prime time slot." Once there was no Thursday at 9 p.m. to win, much of what was built around that slot shifted. "A series no longer had to attract 15 million people on the same night." Seasons could run six or ten episodes instead of 22. A platform could commission a wider slate and learn from what people actually watched, instead of trying to pick a few winners in advance.

None of that shows up if you measure streaming as a cheaper way to deliver the same shows. Hollywood, he argues, first treated streaming as just a new distribution channel. His warning is aimed at today's executives: many companies, he writes, are "making the same mistake with AI today," treating it as "a productivity layer for the existing firm." He calls the link between scarcity and strategy "one of the central ideas" of Reshuffle: "Every generation of strategy is built around managing what is scarce."

A bank that hired people who sell

Customers Bancorp, a banking company with about $27 billion in assets, said in July that a pilot of its "new AI-powered loan closing process" had closed selected commercial loans "in seven days, down from 30 to 60 days typically." In its October results it said it had "hired 50 revenue producers year-to-date while our total headcount trended lower": bankers who bring in new business, added while the rest of the staff shrank. The bank hasn't drawn the line between the two, but together they hint at the possibility this piece is about: when closing a loan stops being the slow part, the slow part may become finding the next borrower.

The objection: savings are real, strategy is a guess

Here is the strongest case against all this, and it deserves a hearing. Hours saved can be counted. You can time the old quote and the new one, check whether overtime fell, and hold someone to the number. "What would we do if quoting cost nothing?" is a guess. It invites a slide full of new markets that nobody will ever check. A chief financial officer who approves the counted savings and ignores the speculation is being careful, not timid. Choudary concedes part of it himself: much of business, he writes, can't be tested by trying two versions and comparing.

All of that is right, and it still misses something. The savings are an advantage mainly while you are the only one who has them. When every distributor can quote in 15 minutes, quote speed stops being an advantage and becomes the minimum. The hours you counted then tend to show up as lower prices for customers rather than profit for you. The companies that get disrupted in these shifts, Choudary writes, are often not the ones that fail to adopt the new technology; they are the ones that keep optimizing around "a previously scarce asset that has suddenly become abundant."

And the guess costs less than the objection assumes. Once there was no slot to win, programmers could commission more shows and let viewers choose. When something stops being scarce, testing what to do with it can get cheap too. If quotes are cheap, you can find out whether small customers are worth serving by quoting all of them for a quarter in one region and seeing who buys. The careful move is to count the savings and run one small, dated test of what you would do differently, so the strategy is something you measured instead of something you hoped.

Many companies ration something because it is slow: the custom proposals written only for large accounts, the second opinion saved for big contracts, the analysis run once a year, the customers nobody calls back. An AI business case that doesn't say which of those you'd stop rationing is only a cost line.