The saving reached the customer. That is the system working, and this paper has spent the day looking disappointed about it

Let me put the case nobody else in this edition will make.

A household paid about £2,437 for conveyancing a year ago and pays about £2,438 now, against inflation of 2.6 per cent. That is a real-terms price cut. The firms doing the work say they are earning less per transaction than they were. The technology arrived in eight out of ten of them.

That is competition transmitting a productivity gain to the customer. It is the outcome the textbook predicts, the outcome every politician says they want, and the outcome this paper exists to look for. We found it, and the response has been to hunt for who is secretly benefiting.

Nobody is secretly benefiting. That is what a competitive market looks like when it works. The gain went to the buyer, in an amount too small to notice, which is precisely how price competition has always delivered gains. Nobody announces it because there is nothing to announce. Your groceries got cheaper in real terms for thirty years and Tesco never sent you a letter about it.

On the evidence problem, my colleagues are right that the speed claims come from suppliers, and wrong about what follows. Of course the numbers come from suppliers: they are the only people who have any reason to count. A law firm does not run a controlled trial on its own title checks. It notices work getting done sooner and buys more of the software. The absence of an independent study is not evidence of a conspiracy. It is evidence that nobody has funded one, which is true of almost everything.

And the price data settles it anyway. If AI had done nothing at all, prices would have risen with costs. They did not. Something made the work cheaper to produce, and the competitive response passed it on. You do not need a stopwatch study to infer that from the price series, which is independent, published quarterly, and built from tens of thousands of quotes.

The slower transaction is the weakest thing in this edition. A hundred and twenty-three days from instruction to completion, up eighteen per cent since 2019, in a chain that runs through local authority searches, lender underwriting, management companies and the Land Registry. Blaming automation for that is like blaming a faster kettle for a slow train.

Here is what I would report instead. The junior roles contracted. That part looks real, even if the numbers we have are supplier numbers. A market that delivers a small real-terms saving to customers while removing the first rung of a profession has made a trade, and it is a trade worth arguing about on its own terms.

But it is a trade, not a scandal. The customer got the money. The paper wanted a villain and found a functioning market, and has spent the day sounding faintly let down.

Sometimes the answer to who got the gain is: you did, and it was not very much, and that is the system working exactly as advertised.