The SaaSpocalypse is coming, at least that is what many self-proclaimed AI experts would like you to believe. According to them, you no longer need software or SaaS solutions because you can build them yourself using AI. In an ideal world with unlimited budgets, time, and people, and no need for security or governance, it’s certainly possible. In all other cases, the SaaSpocalypse will never happen. That doesn’t mean there’s no value in building your own software or agents; on the contrary, Salesforce proves with its Headless 360 that there’s a future in it.
“Headless” isn’t a new marketing term from Salesforce; it has been around among developers for years. Simply put, it refers to an application without a user interface, one that can be used and controlled via code by leveraging APIs and, nowadays, MCP (Model Context Protocol), which enables AI agents to communicate with one another.
Salesforce will make more than 60 new MCP tools available, enabling developers to have their software and AI agents communicate with Salesforce applications, exchange data, and perform actions. This is in addition to the APIs already built into the Salesforce platform. The ultimate goal is that every action for which you currently need to log in to Salesforce.com will soon also be executable via an API or MCP.
Using Salesforce without logging in to Salesforce.com
With Headless 360, developers gain full access to all their organization’s data, workflows, and context within Salesforce. They can build an application or agent on any platform, such as Claude Code, Cursor, or Windsurf. It also doesn’t matter where the application runs, whether that’s on AWS, Google, Microsoft, or in a private data center. That application communicates directly with Salesforce via APIs and MCP and can do everything users can do via Salesforce.com.
This gives developers complete freedom to build applications with the user interface their organization needs. All data and actions within the Salesforce platform can be accessed or executed from that custom-built application. You get, so to speak, the car’s chassis, including the engine block, steering wheel, brakes, pedals, and safety systems, but you build the entire body yourself. Tailored to your taste and needs: a small and simple car, or a very large one with all the bells and whistles.
Isn’t building custom applications a SaaSpocalypse?
Now you might be thinking: if I’m going to build those custom applications, can I also replace Salesforce? Wouldn’t that be a SaaSpocalypse? Technically, that’s possible. But then you’d also have to build the entire chassis yourself: the engine, the steering wheel, the brakes, and all the safety systems. That’s a lot more complex than building a user interface on top of proven, existing technology.
Of course, you can also use AI for this. AI can retrieve best practices on what a CRM, ITSM, or HR package should look like. It can generate a database structure for the desired application, build integrations with Microsoft Entra or Okta so users can log in, add APIs and an MCP server, and even check all code against security best practices.
However, this is just the tip of the iceberg. Many SaaS providers have years of experience doing this effectively. It’s not impossible to build it yourself, otherwise, new players wouldn’t be able to enter the market. But for a single organization, it’s a lot of work, extremely complex, and above all, expensive to do right.
Will all SaaS applications continue to exist?
We can also flip the question and ask whether all SaaS applications will continue to exist. The answer is: no, certainly not. There are plenty of SaaS applications that aren’t keeping up in this AI era. Those that don’t have a decent API, where an MCP server isn’t even on the roadmap, and where the board is still debating whether they should do anything with AI at all. They won’t survive this revolution, not because organizations are building their own versions, but because an existing vendor that does innovate or a startup with an AI-based solution will replace them.
Fear drives stock prices
We visit many SaaS providers, and every single one of them is investing millions or billions of euros in AI, depending on their size. The financial results don’t lie: virtually all of these companies are still growing and turning a profit, sometimes more than in previous years. And yet their stock prices are under significant pressure.
This has everything to do with fear and little to do with reality. Financial analysts and shareholders who don’t fully understand the technology are buying into the SaaSpocalypse narrative. They see AI and conclude that existing software will become obsolete, without understanding how complex it is to actually achieve that. The numbers and technology tell a different story, but it’s less sexy than an apocalypse.
It’s a pattern we see with every major technological shift. The market reacts first to sentiment, not to facts. Ultimately, reality always wins out over the narrative, but that does mean that solid SaaS companies with strong numbers are now artificially undervalued due to a message that simply isn’t true.
SaaS of the Future
There won’t be a SaaSpocalypse, but SaaS as we know it today is certainly changing. All software will be equipped with AI and AI agents that communicate via MCP and exchange data and tasks. That means fewer people are needed, and therefore fewer licenses. SaaS providers will have to adapt their revenue models accordingly; what that will look like will become clear in the coming years.
In that light, the introduction of Headless 360 at Salesforce is an interesting move. It directly addresses organizations’ need to build their own applications and agents with AI, without having to develop a complete underlying platform. We’re curious to see if other major SaaS players will follow this strategy.
Also read: Salesforce wants to be the leader in ITSM