Independent Analysis: AI, SaaS, and Revenue Growth Strategy

Most enterprise AI projects hit a wall before they reach production, and the wall is rarely the model. It is the data. Specifically, companies deploying agents discover that no one can give a confident answer to a simple question: what systems are running, where the relevant data lives, and whether that data is current enough to be useful. That gap has a name. It is an enterprise architecture problem.

I spoke with Oliver Kocs, who leads business development at SAMU, the enterprise architecture management tool built by Atoll Technologies. The conversation started with a question about agentic AI and ended up covering data governance, compliance pressure in the European Union, why business users now care about IT architecture, and whether enterprise architecture tools can survive in a market where software categories are colliding. His answers surface a tension that is worth sitting with if you are evaluating tools in this space.

The Agentic AI Problem Is an Architecture Problem

When an AI agent makes a decision, it queries systems. It pulls data from wherever it is pointed. The problem is that in most enterprises, no one has a complete, current picture of all the systems those agents might reach, what data those systems hold, or what the dependencies are. Oliver framed it directly: "You may not be aware of exactly where to start with that process" without a prior data asset inventory.

This is not a new problem. European companies building toward General Data Protection Regulation (GDPR) compliance had to map personal data across their organizations, and that exercise forced many of them to build a data asset inventory for the first time. SAMU was used for exactly that kind of mapping. The difference now is that AI raises the stakes. An agent pointed at the wrong data source, or a stale data source, does not just create a compliance exposure. It makes decisions on bad information at machine speed.

SAMU's core function is building and maintaining a central repository of an organization's enterprise architecture: the business processes, the applications, the underlying IT infrastructure, and the relationships between all of them. What Oliver described, and what the platform supports, is using that repository as the starting point for any AI initiative. Before you decide what data to feed an agent, you need to know what data you actually have and where it sits. That work does not happen automatically, and it does not happen without a tool designed to manage that complexity at scale.

AI Governance Is About to Become Its Own Use Case

Enterprise architecture has historically been organized around three layers: the business layer covering processes, capabilities, and data governance; the application layer; and the technology and IT infrastructure layer. This structure comes from The Open Group Architecture Framework (TOGAF), which Oliver described as the dominant global standard for the discipline. Within that model, data governance has always been part of the business layer.

What Oliver signaled is that AI governance is likely to break out as a separate layer, the way data governance did. "AI governance will be a separate use case, just as data governance is," he said. That is not a feature announcement. It is a prediction about how enterprises will need to think about accountability for AI systems, what models are running, what data they are using, and who is responsible when something goes wrong.

Compliance Pressure Is Arriving in Waves

Oliver is based in Europe, and his days are increasingly consumed by vendor questionnaires from customers preparing for the Digital Operational Resilience Act (DORA), the European Union's regulation on operational resilience for financial services. What he described is a market pattern where compliance requirements force companies to get serious about data and system mapping, and that documentation work then becomes infrastructure for other initiatives including AI.

For U.S.-based organizations, DORA may feel distant, but the regulatory direction is clear. Governance frameworks for AI specifically for children are already moving through European policy channels, and comparable pressure on enterprise AI is likely to follow. Oliver's read is that compliance requirements create the business case for doing the architecture mapping work that AI deployment also needs. Both pressures point to the same solution.

Business Users Are Now in the Room

One shift Oliver described from his own sales experience is that enterprise architecture conversations no longer happen only with chief information officers and chief technology officers. Business analysts, business architects, and operational stakeholders are now active participants. The reason is not complicated. When people hear that AI might change their role, they want to understand what systems support their part of the business. They want to know what is behind the scenes.

SAMU's design addresses this directly. The platform allows a business user who does not have deep IT knowledge to see which systems support their area of responsibility without needing to understand how those systems work. The intent is to bridge the gap between business and IT, which is what enterprise architects have always tried to do, but with the tool accessible enough that the architect is not the only person who can use it.

A Platform That Competes by Replacing Parts of Other Categories

Oliver was candid about SAMU's competitive position in a way that is worth noting. Enterprise architecture is a category with well-established competitors. SAMU can be replaced, and it sometimes replaces other tools. What he described as SAMU's actual advantage is flexibility: the platform's customizable meta-model lets organizations use it for use cases that would typically require separate software, including Configuration Management Database (CMDB) functions, license optimization, and SaaS rationalization. One platform handling five use cases is a real cost argument in a budget environment where software consolidation is a priority.

SoftwareReviews peer data supports the vendor support and responsiveness claim. SAMU carries a Composite Score of 7.9 out of 10, a Net Emotional Footprint of +96, and has been recognized as an Emotional Footprint Champion in 2022, 2023, 2024, and 2025. The top-rated capabilities consistently land in vendor support, ease of customization, and business value created. For a company of SAMU's size competing against larger platform vendors, those scores on the softer dimensions reflect a deliberate bet on customer relationship quality over feature volume.

The M&A use case is where this flexibility shows up most concretely. When what was formerly DuPont Pioneer, now operating as Corteva Agriscience, went through its IT consolidation, SAMU was used to compress a two-year planning timeline down to six months by mapping duplicate systems and facilitating migration planning. That is a business outcome with a number attached to it, which is a different kind of conversation than a feature comparison.

SAMU's Position: System Architecture, Management, Utility

The name itself is an acronym: System Architecture Management Utility. The tool was built in 2003, originally to support a bank merger. It has been developed from operational necessity rather than from a research framework, which Oliver credited as the source of its flexibility. Organizations can begin with one pain point, whether that is GDPR data mapping, SaaS license sprawl, or AI readiness, and expand use from there without changing platforms.

For buyers evaluating enterprise architecture tools, the practical question is not which platform has the most features. It is which platform can serve as the organization's single source of truth for IT and business architecture as that architecture grows more complex, absorbs AI systems, and faces increasing compliance scrutiny. That is the job SAMU is built to do. Whether it is the right tool depends on whether an organization's current architecture practice is mature enough to use it well, or whether the starting-from-scratch use case requires a simpler entry point.

My earlier research note on SAMU's role in mergers and acquisitions, SaaS optimization, and digital transformation is available on the Info-Tech Research Group site for additional context.