The SaaS Apocalypse Is Wrong: How ServiceNow's AI Pivot Changes Everything

The Narrative That Wiped Out Billions

Wall Street has coined a term for its latest fear: the SaaS Apocalypse.

The logic is straightforward. AI replaces the work of 10 employees. If companies don't need 10 employees, they don't need 10 software licenses. If they don't need 10 licenses, companies like ServiceNow lose 90% of their revenue.

This fear has evaporated billions across the software sector. ServiceNow has been cut roughly in half from its highs.

The math sounds right. Fewer employees, fewer licenses. Simple.

But simple math on Wall Street is usually wrong.

ServiceNow Isn't Being Replaced by AI—It's Becoming the AI Platform

Think of AI as a brilliant new employee. It can work 24 hours a day, never takes a break, processes information at inhuman speed. Incredible.

But even the most brilliant employee needs a desk. A computer. A system that tells them which tasks to do, who to report to, and where to file their work.

ServiceNow is that system.

AI doesn't replace ServiceNow. AI needs ServiceNow to do anything useful inside a company. ServiceNow is already woven into the entire operational machinery of the enterprise. It knows the workflows, the approval chains, the data structures.

As CEO Bill McDermott put it: "AI without a workflow platform is like a brain without a body." It sits in a jar. Not very useful.

Now Assist: From Zero to $750 Million

ServiceNow launched its AI product, Now Assist, and the growth has been extraordinary.

Contract value: zero to $750 million. Target by year-end: $1.5 billion.

For context, ServiceNow's core product—the thing that generates the bulk of its revenue—took 20 years to reach this kind of scale. The AI product is getting there at warp speed.

The Pricing Model Shift That Flips the Script

This is where it gets clever.

The old model charged per seat—per employee. A company with 100 employees pays for 100 licenses. That's where the SaaS Apocalypse thesis comes from: fewer employees means fewer licenses means less revenue.

ServiceNow is shifting to a consumption-based model. You pay based on how much AI work gets processed through the platform, not how many humans you employ. Half of new contracts are already on this model.

Here's why that matters: even if a company fires everyone and replaces them with AI agents, those agents will be processing more workflows through ServiceNow, not fewer. The AI agents run on ServiceNow's infrastructure. More AI usage means more revenue for ServiceNow.

Wall Street sees AI as a threat to ServiceNow. ServiceNow sees AI as the biggest growth opportunity in its history. So far, the numbers support ServiceNow's view.

Valuation: What Fear Has Discounted

A year ago, ServiceNow traded above 100x PE. Today it's around 60x trailing. On a forward basis, using next year's earnings estimates, the PE drops to approximately 14.8x.

That's roughly Microsoft's valuation—except ServiceNow is growing revenue faster. A $5 billion share buyback is underway, and the GSA contract has opened the federal government as a customer base.

But caution is warranted. The stock is down ~50% from highs, and selling pressure hasn't fully exhausted itself. Bounces have been met with continued selling, suggesting the stock hasn't found its floor yet.

If ServiceNow can establish a base around $110–120 with clear institutional accumulation, that would mark a higher-probability entry point. The SaaS Apocalypse created a fear discount. If the fundamentals prove the thesis wrong—as the AI product numbers suggest—that discount will eventually close.

FAQ

Q: If AI agents replace human employees, why wouldn't companies just build their own workflow systems?

A: For the same reason companies don't build their own email servers or cloud infrastructure. ServiceNow has spent over two decades integrating into enterprise operations across IT, HR, security, and customer service. Building a comparable system from scratch would take years and cost far more than subscribing to ServiceNow. The 98% retention rate exists precisely because the switching costs are prohibitive.

Q: Is the consumption-based pricing model proven at scale?

A: It's still early. Half of new contracts use consumption pricing, but the full portfolio hasn't transitioned yet. The risk is that consumption-based revenue could be more volatile than predictable per-seat subscriptions. However, if AI usage grows as projected, consumption pricing should generate higher average revenue per customer than the old model.