SaaS is Dead? Far From it... (Photo by Osmancan Gurdogan/Anadolu via Getty Images)
Anadolu via Getty Images
A 48-hour sell-off in February wiped roughly $285 billion from software-as-a-service valuations. Wall Street called it the SaaSpocalypse. The narrative was simple: AI agents will replace software, the per-seat model is dead and the incumbents are finished. I disagree.
Here is an overview of how major SaaS companies reposition themselves. In short: SaaS is not dead. The interface is.
Anyone can now build anything. The next best SaaS application seems to be just “a prompt away.” At the same time, we have now seen several years of empty LLM demos that never made it into production. Why? Theoretically I can vibecode now my own CRM system, an HR system and, yes, even a finance platform. But can I manage it? Who does the cybersecurity? Who handles compliance? How do you scale to millions of customers or app calls. Vibecode is changing how Engineers work, they do not replace the need for scalable, secure software architecture. Just look around in Reddit and discussions on LLM cost. The trap is thinking that building a tool is the same as running a platform. It is not. This is the trap that "SaaS is dead" thinking falls into. It conflates building a tool with running a platform. The two are fundamentally different jobs.
The most forward-thinking SaaS companies are no longer selling you a user interface. They are selling you access to capabilities that an AI agent can call on your behalf. The screen is optional. The capability is not.
I tracked how SaaS companies reacted to the AI threat across five dimensions.
MORE FOR YOU
Tracking 10 SAAS Companies on how they are positioned in face of the AI Agent opportunity
Lutz Finger
In 2024 Salesforce CEO Marc Benioff has mocked Microsoft’s AI tool, Copilot, as “Clippy 2.0.” At the same time he launched "Agentforce" by December 2025 it was clear that developers building their own MCP servers exposing Salesforce data directly to frontier models, bypassing Salesforce's Agentforce entirely. Salesforce responded with Agent Fabric and a governance layer. The point is not whether you like Agentforce. The point is that Salesforce saw the attack vector clearly, named it publicly, and moved.
Shopify was equally direct. CEO Tobi Lütke’s internal memo made AI a prerequisite for every hiring decision and acknowledged plainly that for LLM shoppers, the storefront disappears.
Booking.com sits at the other end. The CEO described LLM-sourced traffic as "small but growing." That framing treats a structural shift as an incremental trend.
Already three years ago I argued that the real competitive edge for AI-driven businesses lies not in the model but in the data moat. The companies that understood that early are now the ones pulling ahead. A few companies have formulated very well how the behavioral data that they accumulated over time puts them ahead.
Sasan Goodarzi, Intuit’s CEO, put it directly in a February 2026 interview: "Data is the most important moat in all of this." Intuit has data that Amazon, Google, and OpenAI do not have: 100 million customers whose tax filings, credit history, and business cash flow all live inside the same platform. Add Intuit Mailchimp and one gets a closed-loop financial + marketing intelligence stack covering both the consumer and the small business. No frontier model has access to that combination.
Spotify has also built a data moat. The new CEO Söderström made this very explicit. In the 2/26 earning call he said:
What that means structurally for Spotify is that we are building a dataset that never existed, which is the data set of language to music, language to podcast and language to books. We've had the song-to-song dataset, but no one had the language to song dataset. And I want to drive on a point here, which is this is a very specific dataset. You may think it is a canonical dataset, meaning there is a factual answer to, for example, what is workout music. There is no factual answer to what is workout music.
MCP stands for Model Context Protocol. It is the basis of the ‘new user’ of SaaS platforms. Five companies lead here. Intuit partnered with Anthropic on MCP integration and signed a multi-year deal to embed inside ChatGPT. HubSpot, Salesforce and DocuSign opened MCP servers. Shopify created their own Agentic Storefronts by default for every store.
While having an MCP server is the fastest concrete move a SaaS company can make right now, it is not a risk-free move. Publishing an MCP endpoint means expanding the attack surface of your platform. Salesforce learned this directly: within weeks of launching Agentforce, researchers disclosed a critical vulnerability. Salesforce acted and built a governance layer on top: Agent Fabric and Trusted Agent Identity. Again it becomes clear SaaS is needed as trusted, governed layer that raw vibecoded alternatives can never be.
Some platforms are not just useful. They are legally recognized actors. No AI agent can vibecode its way around that. DocuSign and Intuit are building on ground that requires regulatory authorization to stand on at all. But the regulatory moat is the hardest category to replicate and the one most companies do not have. Salesforce, HubSpot, Shopify, and Zendesk all score weak here.
The final question: did companies stop defending their interface and start rebuilding their architecture around the agent layer?
Intuit is the furthest along. They rebranded their AI infrastructure as GenOS, a Generative AI Operating System. Specific agents handle accounting, payments, and customer interactions. The platform is not wrapping AI around existing features. It is rebuilding the architecture to treat agents as the primary execution layer.
On the other end of the spectrum might be Zendesk. They are projecting 500 million dollars in AI annual recurring revenue by end of 2026. A strong number and somewhat driven by their smart integration of AI companies like ultimate.ai. But revenue targets, as good as they look, do not answer structural competition.
But let us be clear: this is early days. The scorecard above reflects announcements, pivots, and strategic positioning — not finished transformations. Most of these companies are still mid-execution. The data lakes are not fully connected. The MCP endpoints are in beta. The governance layers are being stress-tested. The stakes are clear. The race is on.
Global SaaS spending is still climbing toward $315 billion. The market did not die. The interface changed. SaaS is not dead. It is becoming more powerful: the capability layer that the next generation of AI works through.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。