When companies are confronted with a new technological reality, they rarely respond immediately by radically restructuring their business. Instead, they initially respond by defining forms, responsibilities, and visibility. This is not a sign of ignorance. Rather, it is a typical pattern of organizational adaptation: what is new is first named, defined, demonstrated, and given a place within the organization before it gradually becomes embedded in processes, products, and decisions.
A look at the early days of the internet illustrates this very clearly.
When companies began establishing an online presence
When large companies developed their first websites in the 1990s and early 2000s, the focus was often not on the product, but on the organization itself. At first, websites served as a digital representation of the company: featuring executive boards, locations, press sections, organizational charts, career information, and language that strongly resembled annual reports.
That wasn’t wrong. It simply reflected the perspective of the time.
The internet was often initially viewed as a new place where one had to establish a presence. Not necessarily as a new dimension that would transform products, sales, service, communication, and value creation in equal measure. Consequently, online presences emerged that expressed less the new logic of the medium than the existing logic of the company.
One could say: Many companies initially did not build a digital version of their services, but rather a digital version of their self-description.
The Website as a Showcase, Not a Business Model
From today’s perspective, this seems almost touching. Banks were working on an online presence long before online banking became the norm. Retailers built web shops that were often treated as side projects from an organizational standpoint. Corporations established internet or e-business units that did not transform the entire company but were, in a sense, responsible for the new sphere on its behalf.
What is crucial here is not so much the technical maturity of that era as the underlying organizational pattern.
The internet initially appeared as something additional. A channel, a field, a visible domain for which one could assign responsibility. Only over time did it become clear that “online” was not separate from the actual business, but rather transformed its very conditions. It wasn’t just a department that had to go online. The business itself gained a new environment.
New technologies are first organized, not immediately understood
This points to a more general observation. Organizations often respond to technological upheavals first by categorizing them, not by integrating them. They create teams, roles, budgets, programs, responsibilities, and visibility. This is not merely a defensive reaction. On the contrary: it is often the necessary first step that makes something new manageable in the first place.
New technologies do not simply emerge as a finished strategy. They must be made socially, organizationally, and linguistically manageable.
That is why the first steps are often not the most elegant. But they are adaptable. A new department, a new program, a new area of responsibility—all of these are organizational forms that companies use to buy themselves time to learn. Not everything about them will remain. But much of it is a transitional structure, not a misstep.
Perhaps that is precisely why it is worth taking a closer look at AI
In the field of AI, too, we are currently seeing how companies are trying to integrate this new technological capability into their organizational structures. Some are establishing centers of excellence, others are formulating guidelines, and still others are making a conscious effort not to isolate AI in specialized departments but to integrate it across the board. Many decision-makers know full well that it would be risky to treat AI merely as a niche topic. That is precisely why they are committed to involving as many functions and teams as possible in AI-related work at an early stage.
This is a key distinction from the oversimplified narrative that companies “don’t understand the transformation.”
Often, awareness of the problem has long been present. The real challenge lies elsewhere: How do you integrate a technology that is potentially relevant everywhere without either dispersing it or neutralizing it within a central unit? How do you prevent blind activism without falling into institutional paralysis? How do you provide direction without narrowing the field’s openness too soon?
Between a Center of Excellence and a Broad-Based Movement
It is precisely in this respect that the current AI phase resembles earlier technological upheavals, without being identical to them. Today, too, we can observe that companies are searching for ways to deal with a development that clearly extends beyond individual departments.
The tensions are clearly visible.
If AI becomes too centralized, it risks becoming a niche topic. If it is fully decentralized too early, a common framework is often lacking. If it is treated solely as an efficiency tool, strategic questions remain underexposed. If it is discussed only strategically, operational experience is lacking.
There is no mistake in this search. It is an expression of a transition.
What the Internet Years Reveal in Retrospect
Looking back at early corporate communication on the web is interesting because it shows how long it can take for a technology to be understood not merely as an add-on, but as a prerequisite. First, it is made visible. Then it is embedded within the organization. After that, parallel worlds often emerge. Only much later does deeper integration into core processes, decision-making frameworks, and product concepts begin.
At some point, the internet was no longer just the “online” realm. It became the backdrop for almost every area.
It is precisely this shift that is historically remarkable. For it does not occur through a single strategic decision. It seeps in. Over years. Sometimes over decades. Only gradually does the perception of what is actually the core and what is merely the form change.
AI could trigger a similar shift
Whether AI will undergo a comparable development remains to be seen. But there is much to suggest that, in this case too, what is emerging is less a new function than a new layer over existing work. Not with the same intensity everywhere, not at the same pace everywhere, but potentially across knowledge work, decision support, communication, analysis, development, and service.
This changes more than just tools. It changes the question of where expertise actually originates within organizations and how it is distributed.
Because when more people gain access to previously specialized services through voice-based systems, entry barriers, preparatory work, and initial implementation stages shift. This does not mean that expertise disappears. On the contrary: it could actually make its value more visible. Because the easier access becomes, the more important the distinction between basic operational capability and robust expertise becomes.
The Real Tension: Access and Expertise
This is precisely where the current debate becomes particularly interesting. AI and prompting give many people faster access to topics, forms of expression, and work outcomes. This presents an enormous opportunity for productivity. At the same time, it raises the question of how organizations can prevent access from being confused with expertise.
This is not a new concern either. Previous waves of digitalization have already shown that new interfaces easily create the impression that complexity has disappeared, even though it has merely been redistributed.
Perhaps this is precisely one of the most important tasks of the coming years: making technologies widely accessible without shortening the paths to true expertise to the point where they become unrecognizable. Not because new tools are problematic, but because organizations can only function well in the long run when ease of use and sound judgment go hand in hand.
What new technologies initially reveal within organizations
Perhaps this is the most productive way to interpret both the early years of the internet and the current AI era: at first, new technologies reveal less about what they are already changing than about how organizations deal with uncertainty. They reveal the language companies use to describe the new, the structures they build first, and where they try to establish control before clarity has emerged.
Organizations learn in public. And often in ways that, in hindsight, seem provisional. The first websites were such forms. Many of today’s AI programs are likely the same. Precisely for this reason, it is worth not dismissing them too hastily as dead ends, but rather interpreting them as transitional figures—as attempts to first make a new dimension comprehensible before it becomes second nature in everyday life.
In the end, this may be the actual connection between the early corporate website and today’s AI initiative: both are not merely technical solutions, but attempts by organizations to understand themselves under new conditions. And it often takes a very long time for representation to actually become internalization.
Thinking about modernizing public digital services? Talk to us.
We combine strategy, implementation, practical training, and quick POCs for measurable results on a powerful platform like dotCMS.
Ready for the next step? Write us. Excited for your project!
