Pricing When Your Competitors Also Have AI

A freelancer quotes $600 for a job that used to take two days and now takes three hours.

She feels slightly guilty about the rate, so on the next job she quotes $350, and wins it.

A competitor with the same tools quotes $220. She drops to $200. Within a year the market rate is $150 and everyone in it works harder than before for less.

Nobody was undercut by a machine. They undercut themselves, using a pricing method that stopped working when the cost of production collapsed.

That method is cost-plus: work out what production costs you, add a margin, quote the total. It was survivable when costs were real. When marginal cost approaches zero for everyone in your market, it is a countdown.

Why Cost-Plus Collapses

Cost-plus needs a floor, and the floor used to be human hours. If a translation took forty hours, nobody could sell it below forty hours of survival wages, so the price had a bottom.

Remove that and the floor becomes the subscription cost every competitor also pays, so the price falls to whatever the least experienced person in the market will accept.

The deeper problem is what cost-plus tells the buyer: you are buying my production time, so the faster I get, the less I am worth. That is the sentence every hourly and per-word rate quietly speaks.

Price against the value of being right, not against the cost of producing.

The Four Things You Can Still Charge For

When output is abundant, buyers still pay for four things, and all four have prices that hold.

1. Outcome: what the buyer wants (the document accepted, the shipment released, the invoice collected), not the artifact produced on the way.
2. Accountability: your name, your contract, your insurance, and your willingness to carry a defined consequence if the work is wrong.
3. Speed with verification: not fast output, which is free, but fast output someone competent has checked and stands behind.
4. Certainty: a fixed price, a fixed date, a defined scope, and no management effort from the buyer.

None of these are correlated with how long the work takes you, which is the point. If your price tracks your production time, every efficiency gain becomes a discount you hand to the customer.

The Cost of Being Wrong

The number that should anchor your price is the buyer's exposure.

Ask what happens if this work is wrong: a delayed shipment, a failed audit, a penalty, a lost customer, a week of an expensive team's time.

Estimate that number conservatively and price between 2 and 10 percent of it for a recurring service, higher when you carry explicit liability.

A worked example. A regional distributor's customs documentation errors cause four delayed shipments a year, each costing about $9,000 in demurrage, rework, and credits, so $36,000 of annual exposure.

A service that removes three of those four delays saves about $27,000. At $1,100 a month, or $13,200 a year, the buyer is ahead by nearly $14,000 and can defend the decision internally.

Price the same work at cost, perhaps four hours a month at $60, and you charge $240 for a service worth $27,000 to the buyer. Same work, same tools, two very different businesses.

Tom's Pricing, Worked

Tom's service from Lesson 17 is review and signed conformity statements for machine-generated equipment documentation.

The wrong version

He prices per word, as always: 40,000 words at $0.04 for review, which is $1,600 for about twelve hours. That is $133 an hour, and someone will quote $0.02 next quarter.

The better version

He prices three packages against the buyer's exposure, which for a delayed release runs $60,000 to $200,000.

Standard review: $4,500 per document set, ten business days, written report, signed statement of conformity, one round of corrections.

Priority review: $6,800, 72 hours, same deliverables, for release-blocked situations where each week of delay has a cost the buyer can name.

Retainer: $2,400 a month covering up to three document sets a quarter and glossary maintenance, with additional sets at $2,900.

The arithmetic

Suppose a mature year, several years in, settles at four retainers ($115,200), six standard reviews ($27,000), and three priority reviews ($20,400): $162,600.

Costs: tools about $4,800, professional indemnity insurance $2,000, accounting and administration $3,000, a part-time terminology assistant at $900 a month for eight months ($7,200), and $3,000 for marketing. Total about $20,000.

Owner earnings before tax: roughly $142,000 on maybe 28 hours a week, against the $38,000 he earned as a producer.

He should assume a slower path, with year one nearer $41,000, but the shape is the point: the same expertise priced against risk rather than words.

Why it holds

A competitor with the same tools produces the same draft. They cannot produce Tom's twenty years, his error log, his insurance, and his signature, and the buyer is shopping for a release they can defend, not for text.

Leo's Pricing, Worked

Leo's offer is exception handling for small freight forwarders: a supervised workflow that reads inbound exceptions, classifies them, drafts responses, and escalates ambiguous ones, reviewed daily.

The wrong version

A software-style fee of $99 a month, because it feels like a small tool and he is nervous about asking for money.

At $99 he needs 60 customers to earn $71,000 before costs, and 60 small customers is a support burden he cannot carry alone.

The better version

Price against the forwarder's cost of a missed exception, which they say is about $2,000 in credits and goodwill, perhaps twice a month.

Setup: $1,500 for configuration, that customer's exception taxonomy, and two weeks of supervised running.

Monthly: $850, including daily review by Leo, a weekly exception report, and a two-hour response commitment during business hours.

Optional outcome term: $250 a month back if more than two exceptions are missed in a quarter, which signals confidence and costs little when the system works.

The arithmetic

Twelve customers at $850 is $10,200 a month, or $122,400 a year, plus $9,000 of setup fees.

Costs: infrastructure and tools $1,400 a month ($16,800), a part-time reviewer for coverage at $1,600 a month ($19,200), insurance and administration $4,000, so roughly $40,000.

Owner earnings before tax: about $91,000 from twelve customers rather than the sixty the $99 plan required, and twelve is a number he can serve well.

The Race to Zero, and How Services Fall Into It

The pattern is recognizable and hard to reverse once entered.

It starts with competing on volume: more posts, more pages, more reports. Volume is the one thing automation made free, so competing on it is competing where there is no floor.

Then comes the unbundled deliverable: the buyer buys pieces rather than outcomes, and comparable things are priced by comparison. Then a rate card, then the annual efficiency negotiation, then procurement treating you as substitutable.

You escape by changing what is being bought. Sell the accepted outcome rather than the artifact, bundle so the unit of purchase is a result on a cadence, carry a consequence a cheaper vendor will not carry, and publish evidence that your error rate differs from theirs.

The only durable escape from price competition is being a different thing, not a cheaper version of the same thing.

Practical Rules for Setting the Number

Quote a price, not a rate: a rate invites comparison of hours, a price invites comparison of outcomes.

Offer three options rather than one, because a single number invites a yes or a no while three invite a choice, and most buyers land in the middle.

Write your guarantee narrowly and keep it: what you check, what you correct free of charge, what you do not cover. Take local professional advice on liability and insurance before committing in a contract, because rules vary by country.

Raise prices on new customers first, in steps, and watch conversion. If every prospect says yes immediately, your price is too low, and never discount for urgency, which is the buyer naming their cost of delay.

What the Three Readers Do

Tom

Tom rewrites his proposal so the first page states the buyer's exposure in their own numbers, and the price appears on page two as a fraction of it.

His first $4,500 quote feels impossible to say out loud, so he practises it twice before the call and then says it without apologizing or filling the silence.

Two clients accept, one negotiates to $3,900 with narrower scope, and one calls it far too expensive and goes elsewhere, which is the correct outcome for a business not selling to everyone.

He also stops quoting per word entirely, because the two pricing models cannot live in one business without the cheaper eating the other.

Maya

Maya is salaried, so her version of pricing is internal.

She stops reporting campaign volume in her quarterly review and starts reporting avoided cost and avoided risk: eleven claim corrections caught before publication, two of which would have required a public retraction.

She frames her bonus request against that number rather than her output, which is Tom's move in a different setting. If you describe yourself by how much you produce, you will be paid like production.

Leo

Leo does the thing that terrifies him and quotes $850 a month to a forwarder who asked for a hundred.

He loses that buyer and keeps the price, which is right: a buyer anchored on $100 is shopping for a tool and will churn in two months.

His next three conversations start from the cost of a missed exception, and two of them sign. The lesson at 24 is worth more than the revenue: the number you say out loud decides which customers you get, and cheap customers cost more to serve.

Worksheet

1. Write your current pricing method in one sentence, and mark whether it is anchored to your time, your cost, or the buyer's outcome.
2. Estimate the buyer's cost of being wrong on one typical job, listing the components and a total.
3. Estimate the buyer's annual exposure for the whole recurring problem.
4. Set a target price between 2 and 10 percent of that exposure, and write the sentence justifying it in the buyer's terms.
5. Build three packages: a standard, a priority version at 40 to 60 percent more, and a retainer with a defined inclusion limit.
6. Write your guarantee in 60 words or fewer: what is covered, what is corrected free, what is excluded.
7. Calculate annual owner earnings at a realistic customer count, listing every cost, and compare with your current income.
8. Name the one thing in your offer a competitor with the same tools cannot copy, and put it in your proposal's first paragraph.
9. Set the date you will raise prices for new customers, and the conversion rate that would make you reconsider.

Common Mistakes

Passing your efficiency gains to the customer

Halve your production time and halve your price, and you have donated the whole benefit of the technology to the buyer while keeping the work. Efficiency gains go to whoever holds the pricing power.

Quoting hours for work whose value is risk

Say "it only takes me four hours" and you have told the buyer the price should be four hours of labour, whatever those hours are worth to them. Describe the deliverable and the guarantee instead.

Competing on volume

Volume is free for everyone now, which makes it the one dimension where competition is unwinnable. If your proposal's strongest claim is "more, faster", you are already in the race to zero.

Guaranteeing what you cannot control

Promising an outcome that depends on the buyer's behaviour or a third party is how a guarantee becomes a liability. Guarantee what is inside your control: correctness, timing, response, and correction at your expense.

Never raising prices

A price set in your first anxious month becomes permanent if you let it, and it quietly caps the business. Raise on new customers first, in increments, and give existing ones notice and a reason.

The RW Finance Perspective

Pricing power is one of the clearest signals of business quality we look at, and it is an economic idea rather than a marketing one.

A company that can raise prices without losing customers has something customers cannot get elsewhere: a brand, a switching cost, a regulatory position, a distribution advantage, or a reputation for being right. That is a moat in practice.

A company that cannot raise prices is a price taker, and price takers earn returns close to their cost of capital, which is another way of saying they create little for their owners.

Studying a business through a Company Page or the Stock Quality Flower, look for margins that hold through cycles, prices that rise at least with inflation, and customers who renew.

Then apply the same test to your own business or job. Can you raise your price, and if not, what would have to be true before you could? The answer is always the same: own something scarce, and make the scarcity visible.

Answering the Part Question

This Part opened with a question: if one person can now do the work of a small firm, what is the actual constraint?

It is not production. Production is the thing that became abundant, and abundance does not pay.

The constraint is problem definition, distribution, verification, and ownership: knowing which problem is worth solving and for whom, reaching those buyers, telling when the output is wrong and standing behind it, and holding a claim on the result rather than renting your effort into someone else's structure.

Everything in Part V follows, because once ownership is the binding constraint, the next question is what to own.

Lesson 20, Stop Renting Your Entire Economic Life, begins the answer.

Key Takeaways

• Cost-plus pricing fails when marginal cost approaches zero for everyone, because the floor under the price disappears.
• Price against the buyer's cost of being wrong, which has not fallen, rather than against your production cost, which has.
• Buyers still pay for outcomes, accountability, verified speed, and certainty, none of which track your production hours.
• A recurring service can be priced at 2 to 10 percent of the buyer's annual exposure, higher with explicit liability.
• Tom priced the same twelve hours at $1,600 by the word or $4,500 as a signed conformity review.
• Leo needed 60 customers at $99 a month or 12 at $850, and twelve is a number one person can serve well.
• The race to zero begins with competing on volume and ends in procurement; you escape by changing what is bought.
• Quote prices rather than rates, offer three options, and never discount for urgency.
• Pricing power is the clearest evidence of a moat, in a public company and in your own small business alike.
• The constraint in this era is problem definition, distribution, verification, and ownership, not production.