The £200k software build sells for £70k now.
AI compressed the cost of building, the saving went to the buyer, the floor is collapsing.
All true but none of it new.
A £10k website in 2006 bought you almost nothing. A £10k website in 2020 bought you a real CMS, real design, something that worked. A £10k website today, built with AI in the loop, buys more than both and costs less to make.
WordPress did this. Squarespace did this. Webflow did this. AI is just the current wave: the hard, expensive thing becomes easy and cheap, and the price the market pays slides down to meet it.
So you get two options.
1. Deliver the same thing for less. Same build, smaller invoice, more people fighting you for it at the bottom. That's the road the doom posts describe.
2. Hold your price and put more in the box. The client pays the £10k they always paid and walks away with more than that used to buy.
I would pick the second one on purpose. Same price, same margin, a lot more delivered.
That £200k build that now goes for £70k? Only if it's the same build. Keep shipping 2026 value instead of 2023 value and the £200k holds.
"Stack the value" is the easy half of that sentence. The hard half is working out what 2026 value actually is.
That's the real job. Not protecting the price but figuring out what the client needs now that they didn't/couldn't have two years ago, and what you're still charging a premium for that's quietly become table stakes.
The floor only collapses if you let your value sit still.