AI won't kill SaaS. It'll make it fatter.

I work in SaaS, and apparently I should be updating my resume. AI is coming for everything. Every vertical will be disrupted. Why would anyone pay for software when they can just vibe code their own solution? I've been hearing this constantly... VCs posting about it, founders chasing where they think the money is going... and AI coding tools are genuinely changing how software gets built. But this idea that it changes when businesses build their own software? Roll to disbelieve.

The buy vs build calculus hasn't changed#

Good CTOs have always known when to buy and when to build. Build what's core to your business. Buy everything else. This was never about capability because plenty of companies could build their own CRM or analytics platform. They chose not to because it wasn't worth the distraction, and writing the software usually isn't even the hard part. I personally learned this the hard way several times early in my career.

AI makes building faster. It doesn't make building smarter. The reasons you didn't want your team maintaining a billing system or an HR platform? Still there. Actually, if AI makes your team ship faster, you really want them shipping things that matter, not reinventing Stripe.

Yes, AI lowers the barrier to build. Used well, it becomes a tool for lowering OpEx, but lowering OpEx isn't everything. The relative advantage of buying from a specialist hasn't changed. You're buying into ongoing expertise and vision and not just point-in-time capabilities. You want the specialists incorporating AI into their products, not your internal team scrambling to keep pace with every model update.

The "just build it" crowd underestimates the actual work of a real product. Things like security reviews, intuitive permission models, support, enhancements. Integrations with everything else in the stack. Managing this stuff takes attention and human accountability. Attention is what's actually in short supply now. Always has been.

Enterprises don't buy software because it's technically impressive. They buy it because someone picks up the phone when it breaks and transmutes their pain into new features. AI doesn't change any of this. If anything, it makes trust more important. When your vendor is using AI, you need even more confidence that someone is accountable for what the AI does. God knows the AI doesn't give a shit.

Low-code tools like Retool and Zapier are the SaaS products that should actually be worried. AI coding is coming for that layer directly. However, even here the pattern holds: companies that don't want to maintain internal tools today won't suddenly love maintaining AI-generated internal tools tomorrow. The shape of the solution changes. The shape of who wants to own it doesn't.

SaaS is about to get wider#

Things will change, though. The old model of "do one thing really well and slot into a company's ecosystem" is worth less than it used to be. If you can't scale by throwing more AI at the same problem (and you can't because the coordination overhead catches up fast), you scale by going broader. More capabilities within the same control plane. Related products that iterate independently but share infrastructure, security, and trust. Bypass the mythical man month by having more products. Smaller companies become Microsoft...but hopefully with better taste.

This is what customers are going to want when every tool has AI capabilities. Nobody wants five different AI control planes. Nobody wants to figure out how their sales AI talks to their marketing AI talks to their support AI. Nobody wants to reconcile security models across a dozen vendors.

The winning SaaS companies will be wider, not deeper. Ecosystems where the whole is worth more than four separate point solutions stitched together with Zapier. CTOs won't buy 5-10 productivity tools. They'll buy into a single productivity control plane that does the job of all of them.

Is this good for startups? Probably not. It favors whoever already has trust and infrastructure (some people call this "distribution advantage"). Is it good for customers? I genuinely don't know. Consolidation tends to mean less innovation at the edges. But I think it's what's coming.

If startups win here, it's because they caught on to the idea of control planes after than incumbents and were able to growth-hack their way into companies. Give them a little taste of good ai-native software. Like that? Have more. Oh, and the more you use the better it all works together. Embrace. Extend.

The way this breaks is if businesses start outsourcing decisions, not just execution. Hire a CFO AI instead of a CFO. That world looks different, and I don't have a good read on it. However, betting against SaaS feels like betting against businesses wanting reliable, supported software they don't have to babysit. It feels like betting that enterprises will suddenly get comfortable vibing their way to mission-critical tools. I'm not seeing that.

The future isn't AI replacing SaaS. It's SaaS that owns a control surface, not just a feature. The companies that figure out how to use AI while still delivering reliability, security, and support—those are the ones that win. That's still SaaS. It's just fatter SaaS.