Is Generative AI Making Custom Software Normal Again?

An interesting side effect of generative AI is that it might be making custom software normal again.

Right at the beginning of commercial computing, software was bespoke by default. It was built for one company, for one purpose, running on expensive hardware, usually to optimise a specific internal process. It was not generic. It was not mass market. It was a competitive advantage.

Then came packaged software. Office suites. Accounting tools. Installed locally and configured by someone vaguely technical in the business. This was a big step forward. Companies no longer needed to build everything themselves. They could buy tools that worked well enough.

Then SaaS changed the game entirely. Suddenly even small, non-technical companies could access sophisticated systems with a credit card and a login. CRM, finance, HR, marketing automation. Enterprise-grade capability became widely accessible.

But with each wave, something subtle happened. The software moved further away from the company using it.

For the last 15 years, most businesses have faced the same trade-off:

• Buy software that mostly works, and adapt your processes to fit it.
• Or build your own, and take on the cost, risk and overhead of an engineering team.

In most cases, the decision was obvious. SaaS was cheaper, faster and safer. And it came with a hidden cost: your business process slowly started to mirror the tool.

At Vorboss, we took a different view. There was a long-held belief that business process should not be defined by the tools available. Instead, we built internal systems to support how we believed the business should operate. If a workflow gave us an edge, we built software around that workflow.

That worked because we had an in-house engineering team. Most companies do not.

But generative AI changes the economics.

The cost of producing tailored software is falling. The barrier to experimentation is collapsing. The gap between “I have an idea” and “I have a working internal tool” is smaller than it has ever been.

So what happens next?

Do we see non-technical founders building their own internal systems?

• A jewellery designer creating a product customiser rather than paying for a rigid ecommerce plugin.
• A plumber stitching together booking logic, WhatsApp automation and invoicing flows that actually match how he works.

Or do we see the return of small, local software firms?

In the early days of the internet, a local sandwich shop would hire a nearby developer to build a website. It was personal. Contextual. Built around that business.

Are we heading back towards “local software for local problems”?

Not global platforms trying to abstract every workflow into a single interface but tailored systems, built close to the business, shaped by its quirks and competitive edges.

Of course, SaaS is probably not going anywhere. The ecosystems are sticky. The switching costs are real. And generic software is still good enough for many use cases.

But AI might be shifting the balance, not because it makes software cheaper but because it makes differentiation cheaper.

And that is a much more interesting change in my opinion.