Provider or payer?

In the AI wave, a service provider stands on the side that earns or on the side that pays. Before any valuation, it must be clear which.

Every week I see companies that are up for sale. Owner-managed service providers, often in the market for decades, with good customers and a successor who is missing. The first question I ask appears in no teaser: does this company earn from the AI wave, or does it just pay the bill?

The answer says more about value than any number in the P&L. Because the P&L shows the past. This question shows whether the margin will still be there in five years.

Two sides of the same wave

Intelligence is becoming a commodity. Almost free, available to everyone. That has been the situation since open models caught up with the leading labs and prices started falling week by week. Electricity did not make any factory rich either. What was built with it did.

For a service provider this means: the cost of delivering its service is falling. The only question is who keeps that gain.

On the provider side stands a company that sells capacity: checked invoices, managed leases, processed claims, maintained equipment. AI makes delivery cheaper. The customer keeps paying for the result. The difference stays in the house.

On the payer side stands a company that sells hours. It buys licences, its people get faster, and at the next quote the customer asks: why does this still cost the same? The licence costs stay, the pricing power goes. Whoever only buys licences produces costs.

How you recognise the side

The difference is not in the model. It lies in two things you cannot download: customer access and change management.

Customer access: whoever holds the customer relationship decides where free intelligence creates value at all. The model has no customers. The service provider does. Check how long customers stay and whether there are framework agreements. And whether the customer buys a result or a day rate.

Change management: the bottleneck is not the technology but the person who is supposed to work differently. Whoever truly converts teams raises the value. Check whether the company has processes that can be rebuilt. Or whether every task sits in the head of a single employee.

Three questions I ask in every review:

• What does the customer pay for: the result or the time?
• Which tasks in the core business can a machine take over today, and who keeps the saving?
• Can a competitor offer the same thing tomorrow with the same tool?

Does the customer buy time, is the saving passed through, and does the competitor use the same tool? Then the company stands on the payer side. No matter how good yesterday's numbers look.

What this does to the valuation

A payer is valued like a company with a shrinking margin, even if the margin is still stable today. Every buyer who asks the AI question factors in the price pressure. A provider is valued like a company with a growth option. Same industry, same revenue, two different prices.

How wide the gap can become is shown by the platforms that buy traditional service providers and rebuild them with AI. I have written about this. The model that General Catalyst shaped with Marc Bhargava targets the rule of 60: 40 percent margin plus 20 percent annual growth. The path there runs through efficiency gains of 10 to 20 percent in operations. Dwelly in property management has acquired six agencies and doubled the EBITDA margin wherever its own technology runs. Crescendo in customer service has automated over 80 percent of customer contacts and reached gross margins of 60 to 65 percent.

These are examples, not a guarantee. But they show the direction. The value does not come from the software. It comes from the company replacing working hours with results and keeping the saving.

What owners can do now

If you want to sell, move your company to the provider side before the sale. That does not take years. It starts with the price list. Where you bill hours today, offer the result. Where an employee checks every case by hand, rebuild the process. The machine takes the standard case, the human decides the exception.

If you want to buy, do not pay for the AI strategy in the pitch. Pay for customer access and for processes that can be rebuilt. The rest you bring yourself.

The bubble that is bursting right now is not AI. It is the belief that you can buy change instead of leading it. For buyers and owners of service providers, the most important question before any valuation is therefore not which tools the company uses. It is which side of the wave it stands on: with those who earn from cheap intelligence, or with those who pay the bill for it.