https://open.substack.com/pub/goswamis/p/ai-disruption-and-the-future-of-saas?r=5zpq7f&utm_campaign=post&utm_medium=web

OpenAI CFO warns of AI disruption and the future of SaaS makers, as the buy vs build debate reshapes software’s fast fashion era.

Most SaaS makers I talk to are already asking the same question: Will AI kill SaaS, or make it stronger?

This week, Sarah Friar — CFO of OpenAI — threw more fuel on that fire. Speaking at Goldman Sachs’ tech conference, she argued that AI is reshaping the buy vs build debate that has defined software for two decades.

Traditionally, companies bought SaaS tools because building custom software was too expensive, too slow, and too developer-intensive. That safety net is disappearing.

Friar’s words:

“It’s not AI that steals your company. It’s someone using AI deeply that disrupts you.”

What This Means for SaaS Makers

1. AI is flattening the economics of software

   With AI coding copilots and agent workflows, teams no longer need to hoard engineers to ship custom apps. A finance department can now spin up its own bots, dashboards, and integrations — without waiting for IT. Friar said OpenAI’s entire finance team is 18% of the size of similar companies precisely because of this.

2. The “fast fashion” analogy is brutal

   Just like fashion retailers collapsed when fast fashion sped up supply chains, SaaS risks a world where companies spin up their own lightweight apps in days — not quarters. Why buy procurement software if AI can auto-generate a custom workflow tailored to your org?

3. SaaS vendors risk being squeezed from both ends

   • AI-native startups can launch faster and cheaper than ever.
   • Enterprises are starting to insource again, this time with AI’s help.
   • Incumbents like Salesforce, HubSpot? Already seeing stock dips because Wall Street is pricing in this risk.

   The Playbook for Founders

   So what do you, as a SaaS founder, do with this?

   • Go deeper, not wider → Generalist SaaS is under attack. Niche, domain-specific SaaS with built-in AI and deep integrations will still win.
   • Build the rails, not just the train → Products that enable customization (platforms, APIs, agent frameworks) become more valuable than rigid apps.
   • Sell outcomes, not tools → If companies can “DIY software,” you need to sell them the results they can’t replicate easily (data network effects, compliance, customer community, workflows powered by proprietary insights).
   • Embrace AI natively → If your roadmap doesn’t have an AI angle yet, you’re already behind.

We’ve seen this movie before. In the dot-com bubble, most web companies died — but a handful like Amazon and Google defined the next two decades.

In SaaS, the same may play out: 90% of tools risk commoditization, but the winners will be the ones who embed AI and position themselves as irreplaceable infrastructure.

So the real question for every SaaS maker today is: Are you building a disposable app… or a platform others will build on?