SaaS isn't dead. But the version you know might be.

Every few months someone declares SaaS dead. Usually it's someone selling the next thing. But there's a real question underneath the hype, and it's worth sitting with.

The classic SaaS model: pay monthly, get a login, use the tool. It was built around one assumption: software does the work faster, but humans still do the work. You pay for Salesforce, but a human still updates the pipeline. You pay for Asana, but a human still moves the cards. The tool organizes. The person executes.

AI agents break that assumption. If an agent can update your CRM, draft the follow-up email, reschedule the meeting, and flag the at-risk deal, what exactly are you paying the SaaS tool for? The interface? The database? Those aren't worth $50 per seat per month.

This is where it gets interesting. I don't think SaaS disappears. I think the pricing does. Per-seat models make no sense when agents do the sitting. What replaces it is probably outcome-based or usage-based: you pay when something gets done, not for the right to do it yourself.

The other shift is that the moat moves. It used to be features and integrations. Now it's data. The SaaS tools that survive will be the ones sitting on proprietary workflows and context that agents need to function. Lose the data advantage, and you're a UI layer that an agent can skip entirely.

So no, SaaS isn't dead. But if you're building a SaaS product today and your entire value is "a nice dashboard humans interact with," I'd be nervous. The question isn't whether your product works. It's whether it still needs a user.