A consulting firm just quantified which software companies are about to get eaten by AI, and the answer is: most of them.
The Summary
- AlixPartners analyzed 500 software companies and built an "AI Disruption Score" showing only 14% have strong defenses against AI displacement
- The two moats that matter: proprietary data and vertical specialization. Everything else is commoditizing fast.
- The timing couldn't be worse: a $40 billion debt wall hits in 2028 just as AI starts eating SaaS revenues.
The Signal
AlixPartners scored 500 software companies on a 1-7 scale measuring AI disruption risk. The score boils down to two questions: Do you own proprietary data that AI needs but can't easily replicate? Are you deeply embedded in a specific vertical with regulated workflows and domain expertise that takes years to build? If you answered no to both, you're in the danger zone. And 25% of companies analyzed are there right now.
The winners are businesses with "proprietary data, systems of context, ecosystem leverage, embedded workflows, and operating in regulated or critical domains". Think healthcare software with patient data moats, fintech platforms with regulatory compliance baked in, or logistics systems managing physical supply chains. The losers are horizontal tools doing generic tasks that ChatGPT or Claude can now handle natively, marketing automation platforms competing with agent-driven campaigns, and anything selling features that became table stakes once foundation models got good enough.
What makes this more than consultant fearmongering is the debt context. Private equity loaded up these software companies with leverage during the zero-rate era. Now they're facing a $40 billion refinancing cliff in 2028 while their revenue models are under siege. Banks will want proof these businesses can survive AI competition. Many won't have good answers.
AlixPartners called this a year ago, before "SaaSpocalypse" was a thing people said out loud. They're not predicting disruption anymore. They're measuring it.
The Implication
If you work in enterprise software, ask yourself: what proprietary data do we actually own, and how vertical-specific is our knowledge? Generic SaaS is about to get squeezed between foundation models on one side and vertical AI agents on the other. The middle is disappearing. Companies with weak moats need to pick a direction fast: go deep into a vertical or become the infrastructure layer that agents build on top of. There's no safe harbor in horizontal feature parity anymore.
Source: Business Insider Tech