Is SaaS dead in the AI age?

Short answer: no. Longer answer: the easy kind of SaaS is in real trouble, and if you build small software products like I do, that difference matters more than the headline.

Why everyone is asking

Early this year, software stocks sold off hard. Investors decided that AI agents could do the work people used to pay per-seat software for, and commentators started calling it the SaaSpocalypse. By mid-year, several writers were pointing out that beaten-down software stocks were rebounding and that AI hadn't killed SaaS after all.

Panic and recovery inside one year tells you the question is too blunt. "SaaS" covers a huge range of products, and they are not all equally exposed.

What is actually under threat

A thin interface on top of a database, sold per seat, is the easy target. An agent can click the buttons for you, or someone can build a rough custom version with AI over a weekend. Per-seat pricing also gets awkward when software does the work and a human no longer sits in the seat.

What isn't

Systems of record, where a business's important data lives. Products people trust with money, compliance and customer data. Software with years of integrations and distribution behind it. Agents still need somewhere reliable to read from and write to, and many established SaaS companies are building agents into their products rather than waiting to be replaced by them.

The moat was never the code. It's the data, the trust and the distribution.

What it means if you build micro SaaS

First, be careful selling seats for something an agent could do. Usage or outcome-based pricing fits the new world better.

Second, pick problems where trust, data or compliance matter, because those are harder to copy with a weekend prototype.

Third, go narrow. Niche workflows that big vendors ignore, and markets like Ghana where global tools are priced and designed for somewhere else, are exactly where a small team can still win.

Fourth, make your product easy for agents to use. A clean API turns the agent wave into distribution instead of a threat.

My take

SaaS isn't dying, it's growing up. Expect fewer lazy products, more pressure on pricing, and more room for small teams who understand one customer deeply. That's a better deal for indie builders than the doom headlines suggest.

Do you think SaaS is dead? Tell me why, or why not, on X at @narayanbona. I read everything.

Bonanza