For the best part of twenty years, the software you use for your job and for your personal life has been provided by big tech SaaS web applications. But the economics of scale are about to implode. AI coding agents are rewriting the rules as we speak.

The categories of our company administration system include one called software. I always have to explain it to my younger colleagues. No, that’s not for our software-as-a-service subscriptions. It’s for software you buy.

Like, on a CD.

A CD?

But we don’t buy software, they retort.

Exactly. Nobody does. They have SaaS subscriptions.

——————————

The economics of scale that made big SaaS profitable are being threatened from within by Agentic AI.

——————————

Way back in the earliest day of computing, only the largest companies, the ones with the deepest pockets, could afford digital automation. And not only because of the huge cost of hardware. It’s said that it could cost in excess of 200k to write a simple ordering function.

In the closing decades of the twentieth century, personal computers and programming languages such as Python democratised digital automation, and of course the final decade brought the internet.

But software was still expensive, but not only that. Using software meant companies and individuals had to force their own processes into the mould of what the software could accomplish. The need for document interchange did the rest: how one company works is how all companies work.

Big SaaS did nothing to change that other than providing a subscription instead of a CD.

Large scale software solutions need companies, and individuals, to have very much the same needs to be able to give shareholders a reasonable return on the investment into high software development costs.

But now, with Agentic AI coding solutions, it looks very much like the cost of software development is going to drop like a ton of bricks.

——————————

The economics of more of the same is about to be replaced by the power of customisation.

——————————

Traditional software depends on us being more or less the same. If we have the same needs and aspirations, we can use the same software.

The customisation panel of a standard SaaS solution would quickly run out of options if we were to demand absolute control.

Here’s some examples of how standard SaaS falls short.

😡 A CRM like Salesforce may offer custom fields and automation rules, but if your sales process involves integrating with legacy systems, you’re stuck.

😤 Trello might work for 80% of teams, but the remaining 20%, those with specialised needs, are left with clunky workarounds.

🤬 SaaS tools are notorious for poor interoperability. Connecting a marketing tool to a custom ERP often requires middleware like Zapier or custom scripts, which break when APIs change.

And we’re not even talking about the “enshittification” of SaaS: raising prices while lowering quality.

On the other hand, look at what Agentic AI can do.

——————————

Agentic AI coding agents are signalling a turn away from generic software solutions, towards hyper-customisation.

——————————

If AI can reduce the cost of custom software development by as much as 90%, as some early adopters report, the economics of "buying software" flip. Why pay for a SaaS subscription when you can own a self-improving system?

Unlike SaaS, where vendors can change terms or sunset features, AI-generated software can be self-hosted and modified indefinitely.

Of course, this shift isn’t without hurdles. And that’s where we at Django Web Studio can help.

We’ve been doing custom software solutions for clients since we started, more than 15 years ago.

And now, using the latest generation of AI-powered coding tools, we see we can help our customers and their users far better. We can propose solutions, not as a slide deck, but as actual working prototypes, which in some cases can be brought to production at the flick of a switch.

The future is custom. We can now help our customers not by having them conform to software but the software conform to them.