This week's most-discussed piece came from a16z: "Good News: AI Will Eat Application Software." The optimism in the headline is a little jarring — software is getting eaten, and that's supposed to be good news? Meanwhile, Salesforce, Adobe, and Intuit have lost 25-30% of their market value in Q1 2026. Markets have started calling it the "SaaSpocalypse."
I'm a PM at a SaaS company. When I read headlines like these, my first instinct is always: Are we in that category?
What's Actually Changing?
Traditional SaaS runs on a simple premise: charge per seat. More users, more revenue. Growth equals user count times ARPU.
AI breaks that equation — in two specific ways:
1. "Tool" is giving way to "workforce." You used to buy accounting software so your accountant could work faster. Now some companies are buying AI that does the accounting itself — no accountant required. It sounds radical, but for certain workflows it's already happening.
2. Per-seat pricing stops making sense. If an AI agent can do the work of five users, why would a customer pay for five seats? The pricing model collapses.
There's a line in the a16z piece that stopped me: "Software that competes for labor budgets will win. Software that competes for tool budgets will lose." As a PM, that sentence hit differently.
What Does This Mean for Vertical SaaS?
I work on clinical management products — dental practice software, general clinic tools, dietitian platforms. My first reaction was: AI can't replace the actual care delivery. Diagnosis, treatment, patient relationships — those are human.
True. But not entirely.
Think about the back-office work at a clinic. Appointment scheduling, reminder messages, financial reports, inventory tracking, insurance submissions. The bulk of this is already automatable — and in some clinics, it's being automated.
The real question is: Does your software look like a "tool" the user operates, or an "employee" that gets things done? If it can behave more like a proactive assistant than a passive interface, the value proposition and pricing model change completely.
Two Paths, Two Futures
The a16z piece argues there are only two paths left for software. From where I sit as a PM:
Path 1 — Vertical AI solution: You own a specific workflow end-to-end. Automation is deep, outcomes are measurable, pricing is outcome-based. Slower to build, harder to copy. Winners win big.
Path 2 — AI-enhanced tool: You keep the existing UX and layer AI on top. Faster to ship, lower risk — but it doesn't protect you from the SaaSpocalypse. The competitive advantage is temporary.
What determines which path you take isn't really a technical decision. It's about what your customers think they're buying from you. Are they buying a tool, or are they buying an outcome?
Three Things PMs Should Do Now
Enough macro thinking. Here are three practical actions:
1. Understand which budget you're in. If your customers pay from an IT or software budget, you're exposed. If they pay from an operational budget tied to outcomes, you're safer — but you need to deliver more.
2. List the repetitive tasks your product enables manually. Find three things your customers do by hand every day using your product that a machine could do instead. Put those on the roadmap — not as "AI features" but as efficiency wins.
3. Revisit your pricing logic. Per-user pricing may be eroding your perceived value. In this climate, "per clinic" or "per completed workflow" models can be more defensible and better reflect the value you actually deliver.
The Bottom Line
Is SaaSpocalypse real? For certain categories, absolutely. But that doesn't mean every SaaS product is doomed. Your job as a PM is to read where your product sits in this shift — before the market reads it for you.
I'm still an optimist. But ignoring the structural change isn't a luxury anyone can afford right now.