The classic model: Storing knowledge instead of letting it flow
Traditional knowledge management operates on a simple principle: experts enter their knowledge, and everyone else searches for and finds it there. This model has two structural weaknesses.
First, it depends on the active participation of a small group of knowledge holders. These individuals typically have little time, even fewer incentives—and often feel as though they are giving away knowledge that gives them personal significance. Maintenance grinds to a halt as soon as the initial enthusiasm fades
Second, knowledge in day-to-day business operations does not primarily arise through explicit documentation. It is embedded in resolved tickets, in emails between colleagues, in meeting notes, and in the decisions made by an experienced service technician on-site. All of this rarely makes its way into the official knowledge base—and is lost as soon as the person leaves the company.
The numbers speak for themselves: According to the 2024 Gartner Digital Worker Survey, 34% of knowledge workers reported spending at least half of their working time searching for the information they need to do their jobs. This is not a failure on the part of individuals, but a systemic problem—and a clear sign that the traditional model has reached its limits.
Demographic change is further exacerbating the situation. When experienced employees retire or leave the company, they take decades of tacit knowledge with them. Companies that fail to address this in a timely manner only notice the gaps once the situation has already become critical.
A New Paradigm: Knowledge as a Flow, Not a Stock
Fluid Knowledge describes a state in which corporate knowledge is not stored in isolated repositories, but flows continuously and contextually between people, systems, and AI—accurate, relevant, and trustworthy. Behind this lies a simple yet far-reaching idea: Knowledge is not a static inventory that needs to be maintained, but rather a dynamic ecosystem that is fed by day-to-day operations.
The key difference: Instead of entering knowledge into a system once and hoping it stays up to date there, every resolved inquiry, every closed ticket, and every decision becomes a potential knowledge building block—provided the right mechanisms are in place to capture, process, and make it available again.
An important principle here: Well-curated domain knowledge is more important than the AI model itself. Without a reliable knowledge base, even the most powerful language models cannot provide robust answers. “Garbage in, garbage out” applies here with particular force.
The urgency is real: According to a Gartner forecast , companies will discontinue around 60% of their AI projects by 2026—not because the technology fails, but because the underlying knowledge base is inadequate. Fluid Knowledge is therefore not an optional improvement. It is the prerequisite for AI investments to bear fruit at all.
The Knowledge Cycle: The Heart of Fluid Knowledge
Fluid Knowledge builds on the fundamental principles of knowledge management—and consistently takes them a step further. The crucial step: Knowledge flows in a closed loop—it is captured, processed, made available, and continuously improved through its use.
AI is the enabler of this cycle—but it is no substitute for human judgment. It reduces manual effort to a manageable level, takes over routine tasks in curation, and makes it realistic for a company to actually keep its knowledge cycle running.
Fluid Knowledge in Practice: Knowledge in the Work Context
A key feature of Fluid Knowledge is contextual delivery. Instead of forcing employees to visit a knowledge portal, relevant knowledge is proactively embedded into the workflow.
This may sound like a minor UX improvement—but it’s a fundamental difference. Searching for knowledge creates friction. Employees who have to switch to a separate system in the middle of a service escalation often don’t do it. Those who see the right information directly in the ticket system use it.
Today, modern AI-powered systems can automatically generate knowledge drafts from tickets, emails, and notes, check for duplicates, flag outdated content, and identify knowledge gaps. The editorial workload is dramatically reduced—without sacrificing quality control.
The Most Common Pitfalls on the Path to Fluid Knowledge
Fluid Knowledge is not merely a technology project. It is an organizational and cultural transformation. Those who overlook this will fail—regardless of the tools they use.
This becomes particularly clear when looking at the current state of AI adoption: According to the McKinsey State of AI 2025, while 88% of companies are already using AI in at least one business function—only about a third of them have truly scaled it. McKinsey refers to this as “AI Theater”: AI is introduced without the necessary organizational foundation in place. The result is impressive pilot projects that lack sustainable value.
The most common mistakes here: Many companies purchase a new system in the hope of solving the silo problem. Without a prior analysis of knowledge flows, without a clear governance structure, and without change management, the next silo emerges—only with a better interface. Added to this is the problem of input quality: AI can accelerate knowledge, but it cannot create it out of thin air. Companies without functioning metadata structures and without clear ownership of content will not experience a miracle with AI, but rather automated chaos. And finally: a lack of feedback loops. A system without usage data does not improve—and thus loses what defines fluid knowledge.
