Marcus Belke
CEO of 2B Advice GmbH, driving innovation in privacy compliance and risk management and leading the development of Ailance, the next-generation compliance platform.
When Microsoft CEO Satya Nadella stated in the 2025 podcast BG2 that classic SaaS applications would collapse in the „agent era,“ the headline was quickly born: SaaS is dead.
Within a few days, this theory was circulating on blogs, social media, and in tech magazines. Many understood it as a radical prediction of a future in which AI agents would make software redundant and companies would no longer need applications, but only intelligent systems that could do everything.
But this interpretation falls short. The actual message is more subtle and more relevant to companies than the provocative headline.
SaaS isn't really dead
Nadella does not describe the end of the SaaS genre, but rather a change in the level of interaction. Many of today's SaaS applications work on the same principle: they offer user interfaces for simple operations on databases, surrounded by business logic and control panels.
When AI agents adopt this logic in the future, i.e., understanding goals, orchestrating systems, and executing tasks, the access point for users will shift. Instead of opening ten individual applications, they will talk to an agent who will take on the task.
However, the platforms below are not disappearing. On the contrary, their importance is growing. Because when everything comes together at the top, it must be clearer than ever at the bottom how data is linked, who makes which decisions, and how processes interact.
In this transition, applications that have grown over the years into unstable spaghetti code are particularly vulnerable, with interwoven dependencies, unclear data structures, and logic that is difficult to understand. Such applications come under pressure because agents cannot work reliably on them. Stable, structured platforms, on the other hand, are gaining in importance.
Why AI needs structure
Modern AI models are impressive. But they have one thing in common: they can only work with what they find. That is the core of the principle. „Garbage in, garbage out“.
When company data is incomplete, contradictory, or unconnected, AI generates plausible formulations, but not reliable results. Many organizations are currently experiencing this: pilot projects deliver confident-sounding answers that are incorrect in detail. The cause is rarely the AI model used. It is the lack of structure.
A simple example illustrates the problem:
An AI may know the term „service provider.“ But without a model that maps how this service provider is connected to processing activities, contracts, risks, or technical measures, its understanding remains superficial. It lacks the context on which real decisions are based.
For AI to work reliably, it needs more than just data: it needs a Organized database of the company.
Why a structured foundation is becoming a key success factor
Researchers and analysts agree: only when data has been modeled, relationships defined, and governance rules clearly established can enterprise AI become reliable. Companies need structures that show how processes, systems, risks, and responsibilities are interconnected.
That's exactly what Ailance does. Instead of scattered documents, it creates a consistent model of the company's reality, a knowledge graph in which processing activities, systems, suppliers, risks, and approvals are precisely linked. AI can not only search this model, but also understand it.
This structure offers a decisive advantage in the age of agents: context.
It not only makes AI more efficient, but also more reliable and explainable.
What AI can suddenly achieve with this structure
Once the business logic has been clearly modeled, the role of AI changes noticeably.
An agent can then not only answer questions, but also recognize connections. Which Processing Which systems are used? Where do certain data categories appear? Where are there patterns that indicate risks?
It can prepare processes, make decisions traceable, and provide answers in audits that currently require considerable effort to produce. In short, it becomes a genuine work tool.
Without this structure, AI remains superficial. It seems impressive until you confront it with specific questions. This is precisely why many early AI initiatives failed when it came to practical implementation.
Shadow AI highlights the problem, Ailance solves it
The use of AI tools by employees in many companies outside of official processes („shadow AI“) is less a Infringement as a symptom: The organization does not provide them with a structured, secure framework for dealing with AI.
Only when it is clearly regulated which data may be used for what purposes, who has access to it, and how changes remain traceable can a reliable basis be established.
Ailance models precisely this basis, so that AI is not a shadow tool, but is embedded in a responsible system. AI thus becomes part of a responsible system that Transparency creates certainty instead of uncertainty.
So is SaaS dead?
SaaS is not dead. But SaaS without structure will not survive the agent era.
The future belongs to platforms that connect knowledge, operationalize governance, and place AI in a context that is understandable to them.
Those who create structure today will gain speed, reliability, and confidence in their actions tomorrow, giving them a decisive advantage in a world where AI is becoming the norm. Companies that continue to rely on isolated tools and established data landscapes risk unreliable results and increasing complexity.
Companies that rely on structured platforms such as Ailance, on the other hand, create a reliable foundation for AI: clear relationships, traceable decisions, and a logical order that agents understand.
Source: Satya Nadella on the BG2 podcast
Marcus Belke is CEO of 2B Advice and a lawyer and IT expert for Data protection and digital Compliance. He writes regularly about AI governance, GDPR compliance and risk management. You can find out more about him on his Author profile page.
German 
English 
Italian 
French