The role of IT is shifting from a function that simply takes orders to a function that helps create value. This statement appears in many strategy papers. It only becomes interesting, however, when it is translated into decisions: Who will be responsible for a product in the future? Who will operate the platform on which it runs? And who will verify what an AI agent has just done?

These are precisely the questions we’re currently addressing with CIOs and IT leadership teams. The following text contextualizes this development, identifies the points of friction, and describes how we support organizations through this process.

The Evolution of the IT Organization Across Three Eras and Two Paradigm Shifts

The history of the IT organization can be divided into three eras, each separated by a fundamental shift.

The centralized line organization, which emerged in the 1950s and 1960s, viewed IT as a technology provider that should be as cost-effective as possible. Separated from the business—and often in conflict with it—it was controlled through hierarchy and approvals. Technology silos, application silos, and a clear separation between business demand and IT supply.

The professionalized and agile organization of the 1990s and 2000s introduced standards, architecture, and shared services. Change and operations were separated; infrastructure and application services were bundled; and sourcing was made flexible—ranging from in-house data centers to SaaS. Later, product-oriented teams were introduced: business owners as product owners, DevOps based on the “you build it, you run it” principle, and autonomous cross-functional teams.

The agent-based organization is now emerging. Business and IT are merging into teams with shared responsibility for results. Platforms provide reusable building blocks instead of custom-built individual solutions. AI agents handle execution and routine decisions, while humans steer and monitor. Governance is no longer reviewed as an afterthought but is built into the processes.

In our model, these three eras are broken down into seven stages, ranging from technology silos through industrialized IT and product-oriented IT to platform-based value streams and the AI-driven organization. These stages are not a staircase that one climbs step by step. Most organizations are operating across multiple stages simultaneously, and that is precisely what makes steering this process so challenging.

Where Organizations Actually Stand

Self-assessments regularly exceed the actual status quo. Our CIO study shows why that is.

Nearly all of the companies surveyed have an IT strategy that is closely aligned with their business strategy. This point receives the highest level of agreement overall. But when it comes to implementation, the picture changes: a significant portion of respondents still do not view IT as a strategic enabler, and when asked how quickly innovations are actually put into operation, many see a clear need to catch up. Participants cite budget constraints, resource bottlenecks, and a lack of standardization as the biggest obstacles.

Two findings are particularly revealing when it comes to the question of organizational structure. First, the model of collaboration between business and IT has made it into the top three highest-priority areas for action, alongside cybersecurity and IT modernization. The convergence of business and IT is thus no longer a topic for the future but is already on the desks of IT leadership today. Second, the classic line organization still dominates, while agile and product-oriented structures are gaining ground but rarely provide comprehensive support.

Three forces are at play simultaneously

What truly challenges leadership teams today is not a single change. It is the simultaneity of these changes.

• The merger of business and IT dissolves the old separation between supply and demand.
  Product development and operations are moving into the business units, while IT provides the technological and architectural foundation. Widespread dissatisfaction with the quality of IT services is further accelerating this shift. Generative AI acts as an amplifier here because it brings business and technical work closer together, often initially in virtual roles alongside the line organization.
• Platform-oriented value streams are competing with cost pressures for the same budget.
  IT organizations that merely provide low-cost standard services are losing their raison d’être. At the same time, cost pressures are disproportionately affecting IT budgets. Investment in platforms and cost-saving targets must therefore be pursued in parallel, while cybersecurity and digital sovereignty are becoming issues for senior management. Control now matters just as much as efficiency.
• The agent-based workforce is shifting responsibility back to the operational teams.
  In most organizations, AI is still at the assistant level, accelerating existing work. Some are already going further and shifting humans into a steering and oversight role, while agents take over routine execution and routine decisions. This shift will continue until entire ecosystems of agents are operating, with clearly defined points at which a human intervenes. This requires robust data quality, effective governance, and data-driven decisions.

What’s Really Disrupting Operations

The most significant disruptions currently stem from forces that are reshaping management, the division of labor, responsibility, and control.

In doing so, we follow the “Human in the Lead” principle: humans set goals, priorities, and boundaries. At defined critical points, they approve AI results and make decisions where it counts. This requires assessment criteria to continuously evaluate an agent’s reliability—through ongoing evaluation, operational monitoring for drift and misconduct, and regression testing with every model or prompt change.

Whether AI deployment can be scaled in a controlled manner depends largely on a shared platform. Platform engineering consolidates access to models and AI services, establishes uniform guidelines for data, and ensures cost transparency and traceability. Without this foundation, each project builds its own system, remaining an isolated case, and new use cases must start from scratch time and again.

The EU AI Act makes AI governance mandatory and should be integrated into existing compliance and governance systems. Roles, AI inventories, policies, controls, and documented approvals belong exactly where quality and risk are managed today anyway.

The fusion of business and technology is not a new idea. AI and platform engineering are giving it new momentum.

The deeper AI penetrates core processes, the greater the dependence on the provider—with risks extending beyond traditional sourcing: models may be discontinued or altered by updates, results become difficult to reproduce, and decisions must be made regarding which data leaves the organization and under whose jurisdiction the operation falls. The ability to exit and data sovereignty are becoming key management issues.

Two target models that should be distinguished

The platform-oriented organization groups its platforms into three areas. Customer Journey Platforms deliver customer experiences based on reusable code. Business Capability Platforms bundle domain-specific solutions in a modular way and are managed like separate businesses. Core IT Platforms provide cloud, data, and automation. Above this lies a governing body that sets standards, allocates resources, and connects vertical and horizontal platforms. The platform is not an end in itself, but rather the lever that enables decentralized teams to quickly create value.

The AI-driven autonomous IT organization builds on this. Core platforms are planned, executed, tested, and further developed by agent-based systems. People set goals, guidelines, and governance rules. In practice, this development unfolds in three phases: First, each person works with an assistant. Then agents join teams as digital colleagues and take on defined tasks. Finally, humans set the direction, while agents execute processes and workflows and seek confirmation when necessary.

The order is crucial. Those who skip the platform and the data foundation will not end up with autonomous IT, but rather with distributed pilots operating without control.

What this means for your next decision

Based on our project work, we can identify four key points that apply to nearly every organization.

• First: Assess your progress separately on all three fronts. The dismantling of line-of-business logic, the development of platform logic, and the preparation of agent-based operating models proceed at different paces. A blanket maturity score obscures precisely the gap that matters most.
• Second: Clarify responsibilities before rolling out technology. Who is responsible for AI governance and the model lifecycle, and what is that person’s title on the executive committee?
• Third: Treat the data foundation as a prerequisite. For agents, clean data quality isn’t enough—they need accessible systems with clear interfaces, access to the current process status, and documented permissions.
• Fourth: Make conscious decisions about sovereignty and the ability to exit. The deeper AI penetrates core processes, the more expensive subsequent corrections become.

How We Support You in This Process

We work with CIOs, IT leadership, and those responsible for AI to assess exactly where you stand. We bring two key resources to the table: our white paper Are CIOs Becoming Obsolete?, which contextualizes developments related to the CIO role and describes three possible future scenarios, and our CIO Study, which shows how IT decision-makers actually assess the status quo. Both are available for download.

Three formats build on this:

• IT Evolution: The Impulse is a free presentation for strategy retreats, executive meetings, and internal kickoffs. It introduces the stages of development and the target scenarios and sparks discussion within the leadership team. It’s not a sales pitch, but a conversation starter.

• IT Evolution: The Assessment is a half-day workshop with the CIO and IT leadership team, held on-site or remotely. Over the course of four hours, we determine where your organization stands on the development path, which of the three target scenarios from the white paper fits your business context, and which decisions can no longer be postponed.
• IT Evolution: The Check is a customized analysis conducted over three to four weeks for organizations facing a fundamental realignment—such as a CIO succession, a merger, or growing pressure from the business. You’ll receive an executive-level assessment of your operating model, a well-reasoned recommendation for the target state, and a roadmap with specific decision points.

We don’t stop at the analysis. As change makers, we consider IT, business, and organization holistically and support you from assessing your current state all the way through to firmly established implementation. Honest sparring at the executive level, concrete support within teams, and a drive toward the goal until the change takes hold.

The best way to determine which format suits your situation is through a conversation. We’re here to serve as your point of contact. Reach out to us—we’d be happy to explore this further with you.