Harvard Business Review LogoMay 27, 2026Flavio Coelho/Getty ImagesFor two decades, software-as-a-service (SaaS) grew by digitizing workflows. Customer relationship management systems recorded sales activity. Field-service platforms scheduled jobs, retrieved customer histories, and fed upsell opportunities back into sales teams. The model created value by taking opaque or messy information and putting it in workers’ hands. Before AI coding tools, few buyers were positioned to build comparable feature sets in-house.
AI’s Impact on SaaS Will Be Uneven. Here’s What Leaders Need to Know.
The “SaaSpocalypse,” or early-2026 software selloff, has reinforced a tempting through-line for leaders: If AI can build or automate software, most SaaS should be rebuilt internally or shopped to a new vendor with a cost structure that reflects lower building costs with AI tools. But executives are overgeneralizing AI’s impact. Generative tools lower the cost of building deterministic, internally focused software, making many workflow and record‑lookup systems ripe for consolidation or in‑house replacement. But SaaS that combines pooled, proprietary data with predictive tasks is structurally different. These “operational intelligence” vendors embed judgment, not just documentation, capturing long‑tail edge cases that single‑firm data and frontier models cannot easily replicate. Their economics are anchored in avoided risk and expert labor, not license fees. Leaders should map their portfolios against a matrix: are SaaS conducting deterministic or predictive tasks and powered by internal or pooled context, to decide what to rebuild, renegotiate, or double down on today’s realities.
68 words~1 min read
