AI isn't going to kill SAAS software business models

AI is making software creation dramatically cheaper, but that shift won’t automatically kill SaaS business models—especially for established enterprise vendors. The core reason is that SaaS companies don’t primarily profit from writing code; they profit from solving hard, sticky workflow problems and from distribution advantages that keep customers from switching. Even if someone can generate code by typing in plain English, that doesn’t erase the practical barriers that matter in enterprise software: switching costs, entrenched contracts, and the trust built through years of service.

The stock market’s continued high valuation of companies like Salesforce fits this logic. While AI lowers the technical cost of building software, it doesn’t eliminate the commercial advantages that come from existing customer relationships and brand recognition. Distribution advantages—sales relationships, long-standing deployments, and reputations for reliability—create inertia. Customers may debate pricing, but they generally don’t rip out mission-critical systems unless there’s a compelling reason. That inertia is reinforced by how enterprise buyers behave: once a purchase relationship forms around a trusted provider, it tends to persist.

The transcript draws an analogy to Amazon and to Jeff Bezos’s early strategy: the internet changes how people discover and buy, but there’s a limited window when customer preferences are “rubbery.” Companies that capture that relationship early can build loyalty through service and reliability. In the same way, SaaS vendors that already sit inside enterprise workflows are unlikely to vanish just because AI makes new tools easier to generate.

AI also doesn’t stand still for incumbents. Salesforce and other SaaS firms are investing heavily in AI, including product branding such as “Einstein.” The argument is that incumbents have both the motivation and the resources to join an “AI arms race,” rather than being left behind by AI-native startups.

A historical parallel is offered from the cloud era. Adobe faced internal resistance when it moved from licensed software toward subscription cloud-based SaaS. That reinvention was painful, but it ultimately remade the company’s revenue and profit outlook. The implication: enterprise software companies can adapt to AI the same way they adapted to cloud—by shifting their product and delivery model rather than assuming the old model will be replaced.

For startups and solopreneurs, the outlook is mixed. It’s arguably a better time than ever for individuals to build and ship software quickly using AI, with smaller teams and faster iteration. But startups still face the same fundamental challenge: competing against incumbents with distribution advantages. The transcript’s prescription is to double down on AI not just to build faster, but to differentiate—because customers will demand more quality.

The final emphasis is on a cascading effect: AI raises expectations. What used to be a “final draft” becomes the starting point, and that mindset spreads across value chains. As customers expect higher quality from everyone downstream, SaaS vendors—incumbent and startup alike—must deliver more. Net result: SaaS isn’t portrayed as dead; it’s portrayed as positioned to win if companies can leverage AI while defending distribution and solving sticky workflow problems. The speaker ends by inviting disagreement, but the thrust is clear: SaaS margins and revenue prospects remain intact, and the market’s valuations are broadly justified.