AI Isn't Killing SaaS; the People Who Sold It Are

As hundreds of billions of dollars in software market value evaporate, a convenient villain has emerged: AI is killing software-as-a-service (SaaS). There’s only one problem: The “SaaSpocalypse” narrative is wrong.

AI isn’t killing SaaS. The people who sold enterprise software are. I know, because for 20 years, I was one of them.

How the Overselling Machine Worked

Enterprise software sales have followed a simple playbook: sell customers on buying for the company they hoped to become, not the company they actually were.

Customers would tell us they needed a certain amount of software. We’d look at their growth projections (which were always optimistic) and build a three-year plan around those assumptions. Buy this much capacity now, we’d say, and the per-unit rate would be dramatically lower. That was usually enough to get them to sign up.

But reality rarely matched the slide deck. Customers bought future-state capacity they’d never use, but the rate was too good, and the contract was done.

Over time, this approach effectively became the industry standard. Salesforce and Oracle did it. SAP increasingly pushed bundles that included customer experience and analytics, focusing on migrating its existing install base to a new version rather than winning new customers.

As a result, companies today are sitting on enormous quantities of software they bought but never used. That’s what’s coming back to bite these SaaS companies.

AI Broke the Model

As AI-driven workflows let smaller teams handle what once required a much larger headcount, the traditional per-seat licensing model comes under real pressure. Years of aggressive overselling have turned into a visible surplus that many CFOs are no longer willing to justify.

The response from software vendors has been revealing, as they’ve rushed to incorporate AI into their products. But do these updates represent new capabilities, or are they just repackaging that existing unused capacity?

When a company offers to swap thousands of unused CRM licenses for seats on a new AI product, that’s not necessarily adoption. In many cases, it’s just converting one line of dormant shelfware into another under the guise of “innovation” to protect the contract value.

Vibe Coding Won’t Save You

Some finance and technology leaders are looking for ways to bypass vendors entirely. The emergence of AI-assisted development, or vibe coding, has made it possible to prototype a functional application in hours instead of weeks or months.

But an application you can prototype in an afternoon still has to be able to stand up to testing, security, maintenance, and support requirements.

Vibe coding can feed a culture of “shadow IT,” where teams build and run tools without the knowledge or approval of the company’s IT department. It may be a fast way to create, but it can create a burden that ends up costing more than the SaaS licenses you were trying to avoid. It’s trading visible waste for hidden debt.

What Change Actually Looks Like

I know all this because I’ve seen what happens when you do it differently. After a career spent perfecting the overselling playbook, I joined a company where we had to rethink the model from scratch.

Instead of charging for how many people log in to our software, we price based on the revenue a customer manages through our platform. What we earn is directly tied to the value we help deliver. If the customer doesn’t grow into the tool, we feel it, too.

When your pricing is tied to outcomes, you can’t afford a 12-month implementation followed by years of shelfware. You need to show measurable impact in weeks, not quarters.

It also forces you to confront an uncomfortable truth about the market. Most companies that could benefit from pricing software aren’t using any. Industry estimates suggests that of the more than 20,000 companies that could benefit from price management solutions, only about 1,200 use one.

The old model oversold the top of the market and ignored this massive middle, because selling outcomes to a $200 million distributor wasn’t as lucrative as loading up a Fortune 500 company with seats they’d never use.

That’s starting to change. It’s an uncomfortable shift for many companies, but it’s the correction the industry needs. The companies that get there first will earn the trust of a new wave of customers.

“SaaSpocalypse” makes for a good headline, but it isn’t quite accurate. The real story is that the bill is coming due for how software was sold. The sooner both sides of the table acknowledge that, the sooner we can build something that actually works.

Bob Pedersen is chief revenue officer at Zilliant, a provider of pricing life cycle management. A former VP at Salesforce, Pedersen has more than 25 years of experience managing high-performing sales teams.