Will AI replace SaaS, or simply make every SaaS product more personalized?

AI is making custom software significantly easier to build.

But easier to build doesn't automatically mean easier to own.

For years, the build vs buy decision was relatively straightforward.

If a capability wasn't core to the business, buying SaaS was usually the better option.

You got a mature product, regular updates, integrations, security and support without building an engineering team around it.

AI changes part of that equation.

A small team can now build internal tools and business-specific workflows much faster.

But there is another side to the equation.
Who maintains it?
Who handles security and infrastructure?
Who monitors it in production?
Who upgrades it when dependencies change?
Who owns it when the internal tool becomes critical to the business?

This is where I think the real shift will happen.

Companies won't stop buying SaaS.

They will become more selective about what they buy and what they build.

Generic capabilities will continue to make sense as SaaS.

Business-specific workflows may increasingly become custom software.

AI reduces the cost of building software.

It doesn't remove the cost of owning it.

So the important question for engineering leaders won't be:
"Can we build this?"

It will be:
"Is this important enough to own?"

That distinction could shape how companies build their software stack over the next few years.