SaaS is losing its grip — then what can we expect to happen?

Because of AI, companies can now code their own custom software solutions in record time.

The interesting question isn't whether the SaaS space will get smaller. But what will happend once companies start to build their own software

AI is making it faster and cheaper to build software. That's real. Companies that used to need a 6-month procurement process and a €50k/year subscription can now spin up internal tools in weeks.

As David Heinemeier Hansson (DHH) recently put it on the Danish podcast SaaS Købmænd: major US SaaS companies are being valued as if they'll be dead within two years. The market is repricing the model in real time.

So yes - SaaS is losing some of its grip. Not overnight. But steadily.

And I get the appeal. I vibe code apps myself. The speed is intoxicating. The feeling of building something that fits your exact workflow instead of adapting to someone else's product? Hard to beat.

But here's what I think most people underestimate: the hard part was never building the software. The hard part was the clarity behind it.

If you are not building with a product mindset then the product is a product of your own biased beliefs, and very few are activly aware of this.

After 25 years of building software for enterprise, one thing is clear to me: as long as people are the ones driving the decisions, the output will reflect the perspective of whoever is in the room. If that perspective is narrow — and especially if it's siloed in IT — the product will be too.

A big part of those 25 years has been spent listening to leaders and management who have a clear but biased opinion about what they think they need. Asking the right clarifying questions. Challenging assumptions before they become features. Building clarity before building code. Because no one — regardless of how much coding power or budget they have — knows what to build unless they're properly equipped with a product mindset.

Code Looks Like a Product. But Code Is Not a Product.

A product is shaped by everything that happens around the code. The field studies. The stakeholder meetings. The user interviews. The compliance reviews. The security audits. The prioritization debates. The design critiques. The failed prototypes.

And perhaps most importantly - the hard conversations about what the software should not be. What to leave out. What to say no to. That discipline is what keeps a product focused instead of bloated. It requires a keen product mindset, and without it, you fall straight into what Melissa Perri calls the build trap: shipping more and more features without solving real problems.

All of that is invisible in the final interface. But it's what creates clarity — clarity about who the product is for, what it should do, what it should not do, and why. And that clarity is the reason the product is secure, compliant, performant, and actually usable.

When you replace a mature SaaS product, you're not just replacing code and features. You're replacing that entire layer of context - years of product thinking baked into every decision.

What Actually Makes Up a Product

If you look at what actually makes up a mature product, the code is just the center. Around it sits product management, product design, problem discovery, solution validation, security, compliance, governance, and operations.

Vibe coding covers the center: the engineering. Everything else is what separates code from product - and skipping it is exactly how you end up in Melissa Perri's build trap.

AI Can Build. But It Can't See.

AI can generate interfaces, write code, and structure flows. It can even suggest security principles and help with documentation.

But AI doesn't create clarity. Not the kind that matters.

It doesn't observe your users in the field. It doesn't understand adoption friction. It doesn't navigate organizational politics. It doesn't know your local compliance landscape. It doesn't usability-test itself.

The Pain Doesn't Show Up in Week One

And here's the tricky part - it won't hurt at first.

The first reaction is almost always positive. Leadership sees savings. Something gets shipped fast. The team feels momentum.

The pain shows up in month 8. Or year 2.

Too many clicks. Weird workflows. Security gaps. Compliance risks. Users quietly working around the tool instead of with it.

This doesn't mean "don't build your own." It means: don't build blind.

Specialists Become the Snipers

Last summer, a team lead at one of Denmark's largest enterprise companies asked me: what happens to specialists when generalists become the AI pilots?

My answer: specialists become the snipers.

They're the ones brought in to create clarity - not just for the software, but for the AI agents building it. They steer both the product and the prompts in the right direction for the company.

Because here's the thing. If management wants to save money by building in-house for a more custom fit, they need to think hard about the echo chamber the AI pilot is operating in. And that pilot is often a software engineer sitting in IT - capable, but siloed. Building from a narrow slice of the full picture.

Sure, AI tools can accelerate the meta layer — research, prioritization, validation, security, compliance, product thinking in general. But I don't see AI owning all of that any time soon. Not even with an AI-first approach.

The Post-SaaS World Won't Lack Code

I've written before about how the real bottleneck in product organizations is no longer production capacity - it's clarity at speed. AI has shifted the constraint. We can build faster than ever. But building fast without clarity about users, needs, priorities, and constraints just means we produce mediocre solutions to no problems - at scale.

That's the exact trap waiting for companies replacing SaaS with their own AI-built tools. The production is easy. The clarity is hard. And clarity is what mature SaaS products spent years building into their operating model.

The companies that will win in a post-SaaS world aren't the ones that write the most code. They're the ones that wrap fast AI-assisted production in real product thinking - user insight, security, compliance, governance, and continuous improvement.

There's a new role emerging here. Not the old-school consultancy that bills for massive teams and long timelines. Something leaner. Someone who helps companies build fast - but build right. Who brings UX, security, compliance, and product mindset to the table when the code is already flowing.

Because the post-SaaS world won't lack code.

It will lack clarity. And without clarity, there is no product maturity.

And if companies think they can replace SaaS with a vibe-coded app, they're confusing code with product.

I work as a UX/Product Design Lead and spend my days doing observational studies with real users. The gap between what gets built and what actually works in context is something I see up close — every single week.