The Price Of Cheap

A client asked me a question last month that I did not have a clean answer for, and I have been chewing on it ever since.

She said, not unkindly, isn't the AI doing most of this now?

She was not trying to negotiate. She was genuinely curious, and she was mostly right, and the awkward pause before I answered was the sound of me realizing I had never worked out what I actually believed about it.

You are going to get this question. If you have not yet, you will inside a year, because the cost of production has become public information. Every operator now knows roughly what a thousand words costs to generate and roughly what a design iteration costs and roughly what it takes to summarize a hundred documents. The prices are on a webpage. Your client can read the same webpage you can.

Here is the uncomfortable structural point, and it applies whether you sell services, consulting, agency work, or anything scoped by effort.

If your price is anchored to how long the work takes, then every efficiency gain you make is a revenue cut you inflict on yourself. You get faster. The job takes six hours instead of twenty. Your invoice, if you are honest with your timesheet, drops by seventy percent. You have built a machine that punishes you precisely in proportion to how good you get at your job.

Most people respond by quietly not being honest with the timesheet, which works for about a year and then stops working, because eventually the client does the arithmetic themselves. And when they do, the conversation is not about price. It is about trust, which is a considerably worse conversation to be having.

The tool prices are still falling, by the way. A capable model aimed squarely at business automation now runs well under a dollar per million tokens of input. That floor is going to keep dropping for at least another few years. Any pricing model that depends on production staying expensive is a model with a countdown on it.

Hold the price and hope nobody asks. This is what most people are doing right now. It is not a strategy, it is a delay, and the cost of the delay is that when the question finally comes you answer it defensively instead of from a position you prepared. Defensive answers lose accounts.

Cut the price to match the new cost. This feels honest and it is a race into a wall. The floor is heading toward zero. You cannot win a competition on a number that is falling faster than you can cut, and every competitor with less overhead than you will get there first. Also, the moment you reprice on production cost you have taught your client that production cost is the basis of the relationship, which means you will have this conversation again every single time the tools improve.

Reprice on what is actually scarce. This is the only one that survives, and it requires you to be specific about something most people have never had to articulate, which is what exactly you are being paid for.

Sit down with a blank page today. Not tomorrow, today, because this takes twenty five minutes and it changes how you answer the question when it comes.

Column one: what I do that is now cheap. Be brutal. Drafting. First pass research. Formatting. Summarizing. Producing three variations of a thing. Transcribing. The initial version of almost any deliverable. If a competent person with a good tool could get to eighty percent of it in an hour, it goes in column one.

This column is going to be longer than you want it to be and that is the point of the exercise.

Column two: what got more valuable because column one got cheap. This is the column nobody writes and it is where the money is.

When producing options costs nothing, choosing between them becomes the bottleneck and therefore the valuable act. When anybody can generate a strategy document, knowing which strategy survives contact with this particular market becomes the scarce input. When drafts are free, taste is what you are selling, and taste does not have a price per million tokens.

Editing got more valuable, not less. Judgment got more valuable. Saying no to four of the five proposals got dramatically more valuable, because now there are fifty proposals.

Column three: what nobody else can do at all. Your relationships. Your specific pattern recognition from having watched this industry for eleven years. The fact that you have seen this exact failure before and know how it ends. Being the person who is accountable when it goes wrong at two in the afternoon on a Friday.

That last one deserves more weight than people give it. A large part of what clients buy is somebody to be responsible. A tool cannot hold a consequence. You can, and that has always been half of the actual product even when nobody wrote it on the invoice.

Stop selling hours. If you are still billing time, you are selling the exact thing that is deflating. Move to fixed scope with a defined outcome. The transition is uncomfortable for about two clients and then it is simply how you work.

Price the decision, not the document. The report is now nearly free to produce and both of you know it. What is not free is the two hours of your attention that determined which three questions the report should answer. Restructure the offer so that the thinking is the visible line item and the artifact is the byproduct.

Add the tier that used to be uneconomic. This is the opportunity everyone misses while they are busy panicking about the downside. There is almost certainly a service you could not offer profitably before because the production cost was too high for the price your market would bear. Go look at that idea again. The arithmetic has changed. Delivering a productized version through something like Go High Level means the delivery overhead does not eat the margin the way it would have two years ago.

Find out where your hours actually go first. Every one of the moves above depends on knowing which parts of your work collapsed and which did not, and almost nobody knows this accurately. People guess, and the guess is reliably wrong in the same direction. A week of actual measurement with something like Rize will tell you more than a month of thinking about it, and the gap between where you think the time goes and where it goes is usually the whole answer.

When the question comes, and it will come, here is roughly what I say now. Steal it and make it sound like you.

Yes, quite a lot of the production is automated at this point, and honestly that is the reason the turnaround dropped from three weeks to four days. What you are paying me for is not the typing. It is that I have watched about forty of these and I know which two of the six obvious options actually work for a business your size, and I am the one on the hook when we choose wrong. If that stops being worth the money, tell me and we will look at the number together.

Three things are doing the work in that answer. It concedes the true part immediately, which defuses the whole thing. It reframes the purchase without being defensive about it. And it invites the price conversation rather than dodging it, which signals you have thought about this and are not afraid of the question.

The version that fails is the one where you get slightly prickly and start explaining how much skill is still involved. Everybody can hear the anxiety in that answer, and the anxiety is what they price against.

Some of your current work is going away and no amount of repositioning saves it.

If a meaningful share of your revenue comes from clients who are buying production volume at a low rate, that segment is leaving, and the correct response is to let it go on purpose rather than fight for it at declining margin until it leaves anyway. Fighting for the bottom of your market is how people spend two years getting poorer while working harder.

The counterweight is that the top of your market is getting better, not worse. Clients who were previously priced out of your judgment because it came bundled with expensive production can now afford the judgment on its own. That is a real expansion and it is available to anyone willing to unbundle.

The single practical thing I would leave you with is about timing.

There is an enormous difference between repricing because you have thought it through and repricing in response to a client who has just done the arithmetic out loud on a call. The first one reads as confidence. The second reads as getting caught.

You have some runway. Most clients have not run the numbers yet, and the ones who have are still being polite about it. Use the window. Do the three columns this week, restructure one offer this month, and have your answer ready before somebody makes you improvise it.

Because the question is coming. Mine came in a perfectly friendly voice on an ordinary Tuesday, and the pause before I answered cost me more credibility than any number I could have quoted.

Jordan

The AI Newsroom is written for people who run something. Forward it to the one person you know who needs it.