If you look at a typical European SaaS stack today, it’s still a zoo: a tool for sales, another for docs, three for project management, plus a handful of copilots bolted on top. AI usually shows up as a sidecar, a scoring model here, an assistant there, while the core product is still UI plus integrations. That made sense when money was cheap and every marginal productivity tool could justify its own licence.

The constraints are very different now. In 2024, European B2B SaaS funding was still around €4.9bn, one of the largest verticals after climate tech, but that capital is being deployed into fewer, more ambitious platforms. At the same time, the software side of the generative AI market is on track to become a multi-tens-of-billions category by the end of the decade, growing much faster than the rest of enterprise software. CFOs and CIOs are reacting the way engineers do when a codebase gets messy: they refactor. They don’t want 40 tools; they want a small number of surfaces where AI shows up as a native capability, not another browser tab.

In practice, that means most “AI SaaS” ends up in one of two buckets. Either it becomes a feature inside a suite, something that lives as an app, plugin or agent on top of O365, Google Workspace, Salesforce, Notion, etc. or it is one of a small number of vertical platforms that genuinely earn the right to be their own surface because they own workflow, data and outcomes ​end-to-end.​

If you sit in on European budget meetings today, you can feel the shift. Ten years ago, great SaaS meant clean UX, solid integrations and predictable ARR. Intelligence was a bonus. Tools logged activity; humans still did the work. As long as churn was low and NRR hovered above 110%, nobody asked too many questions.

Now, every serious buyer walks in assuming automation, not data entry. In 2024, funding to Europe-based startups was about $51bn, and roughly $11bn of that went into AI companies alone. At the same time, the generative AI software market is projected to grow from around $11bn in 2023 to more than $50bn by 2028. In that world, a CRM that does not move pipeline, an FP&A tool that does not propose the budget and a support suite that cannot resolve tickets autonomously is hard to justify at a growth multiple. The application layer, the workflow and the automation logic collapse into one product, and the value accrues to whoever reliably shifts the metric that matters, not whoever owns the login screen.