We are heading for the SaaSpocalypse as AI upends business models: David Richards

From the end of the meeting to something tangible took about an hour. No committee. No developers. No six-month specification cycle. No design sprint theatre. An hour. A company, ready to talk to customers.

Now imagine you are an enterprise software salesperson. You have flown in, polished your deck, rehearsed your demo. You sit across the table from a prospect and begin your pitch.

Somewhere around slide four, one of them opens a laptop. By slide seven, they have typed a prompt. By the time you reach your pricing page, they have built the thing you are trying to sell them. Not a rough sketch. A working prototype, configured to their own workflows, generated in the room whilst you were still talking.

That scenario is not science fiction. It is this year.

What we are watching is not disruption in the comfortable, MBA-seminar sense of the word. It is something uglier and faster. Call it the SaaSpocalypse: the moment when the economic foundations of an entire industry crack simultaneously, not at the edges but at the centre.

The SaaS model was built on a simple and beautiful idea. Build software once, sell it to thousands of companies on a subscription, collect the seats, harvest the renewals. The moat was the product itself. Proprietary code. Special workflows. Years of accumulated complexity that made switching painful and replication expensive.

Investors priced that moat generously. Valuations reflected decades of assumed recurring revenue.

But what is a seat worth when the software can be rebuilt in an afternoon? What is a renewal worth when your customer’s intern can replicate your core product before lunch? What is proprietary complexity worth when AI dissolves it on demand?

My own team has been testing this directly. In roughly an hour, individuals have built a bank, a CRM, an accounting package and a defect-tracking platform. Not enterprise-grade, globally regulated systems. But close enough to functional that the question has shifted from whether this is possible to how long incumbents have before it becomes normal.

The answer, I suspect, is not long.

The seat model dies first. Why pay per user when a single AI agent can do the work of 20? The enterprise contract dies next. Why negotiate a seven-figure renewal when you can rebuild the application over a weekend? And then the valuation logic collapses, because SaaS multiples were always a bet on the durability of the moat, and the moat is draining.

This does not mean every software company disappears overnight. The ones with genuine defensibility survive: deep regulatory integration, proprietary data that cannot be replicated, trust relationships built over years, distribution that no prototype can shortcut.

Salesforce is not dead. But the long tail of SaaS, the thousands of competent point solutions charging 12 seats a month for something Claude Code can assemble in 90 minutes, that is a different conversation.

When I started my first company in Silicon Valley, building a product took months and considerable money. That friction was the price of entry, and it protected those already inside the walls. The friction is gone. The walls with it.

So here is the question that technology businesses need to answer honestly, and soon. If a motivated competitor could rebuild your core product in an afternoon, what exactly are you selling? If the answer is the software itself, you do not have a business. You have a countdown.

David Richards MBE is managing partner of Yorkshire AI Labs