The data platform’s CEO argues that agentic AI doesn’t threaten SaaS infrastructure; it stress-tests it. The tools built on trusted data and rigorous measurement won’t just survive the transition. They’ll define it.
In each installment of SaaSpocalypse Watch, we put one industry voice in the watchtower to offer a frontline view of how agentic AI is reshaping the SaaS landscape. This week, Funnel CEO Fredrik Skantze makes the case that the real question was never whether AI would disrupt the marketing stack, but what it would find when it got there.
There’s a version of the SaaSpocalypse argument that Skantze finds straightforwardly mistaken – and it’s the popular one. The idea that agentic AI will allow companies to build their own marketing workflows and bypass the tools they currently pay for assumes that the tools are the point. They’re not. The infrastructure underneath them is.
“The SaaS tools that will thrive are the ones that become the infrastructure those agents operate on,” he says, “not the ones that just wrap a chat interface around a static dashboard.” Skantze’s view is that AI is an evolution of the SaaS model, not a repudiation of it, but only for the companies that have something worth building on. For those that don’t, it’s a different story.
The most exposed category, in his reading, is point solutions: tools that do one narrow thing and depend on human effort to connect them to everything else. Standalone reporting platforms that don’t own their data layer. Connectors and ETL pipelines that just shuttle data between systems. Basic dashboards with no semantic understanding of what the numbers actually mean. These aren’t just vulnerable to disruption; they’re the products that agentic AI will bypass entirely. And the complexity-as-moat argument doesn’t hold either. If your product requires a consultant to operate and an agent can now do that work, the value proposition doesn’t erode gradually; it collapses.
What remains essential is trust and, specifically, the infrastructure that earns it. Agentic AI is only as good as the data it runs on and fragmented, inconsistently defined or unvalidated data doesn’t produce faster insights; it produces wrong answers faster. The foundation that ensures data is clean, standardized and carries business context becomes more critical as the speed of execution increases, not less. A semantic layer that knows what ROAS means in your specific business context, rather than as a generic industry metric, is the kind of detail that separates reliable agentic output from expensive noise.
There’s a second indispensable element that Skantze identifies and it’s one that tends to get lost in the infrastructure conversation: strategic accountability. The discussion between CMOs and CFOs about whether marketing investment is delivering incremental growth isn’t going anywhere. If anything, AI raises the bar for it because the tools now exist to measure what’s working with genuine rigor. The organizations that have built proper measurement infrastructure will be able to make that case with the same discipline a finance team brings to any capital allocation decision. The ones that haven’t will keep falling back on platform-reported ROAS and intuition and that position is getting harder to defend.
Funnel processes $90bn in ad spend annually across 2,000 advertisers and 1,000 agencies, a data foundation that no individual company could replicate internally. Skantze says the platform has been adapting for the AI era for some time, moving from a reporting tool toward a system that works for customers rather than one customers have to work. Tasks that previously required a data scientist, building a marketing mix model, setting up multi-touch attribution, diagnosing a sudden drop in performance, can now be initiated with a plain language question. The modeling work that used to take weeks is being automated.
Want to go deeper? Ask The Drum
The democratisation argument is where Skantze gets most animated. Sophisticated measurement, triangulating across MMM, multi-touch attribution and incrementality testing, has historically been the preserve of large enterprises with dedicated data science teams. The operational and financial barriers made it inaccessible to anyone without significant resource. Agentic AI dismantles that entirely, making the same analytical capability available to growth-stage companies at a fraction of the cost and complexity. The competitive advantage that came from being able to afford rigorous measurement is about to become a lot harder to sustain.
In five years, Skantze expects the B2B marketing stack to look materially different, with fewer tools, doing substantially more, with the boundaries between data warehouses, measurement platforms and reporting dissolving into something more unified. The sprawl contracts. The infrastructure that remains becomes correspondingly more important.
The warning buried in that outlook is also the opportunity. The tools that will define what marketing technology looks like on the other side of this transition aren’t the most feature-rich or the most visible. They’re the ones with the most trustworthy data underneath.
B2B World Fest, Miami, November 8–11, 2026. Book your place at our festival of B2B insight and B2B awards celebration. To be among the winners, enter the Drum Awards for B2B today.