What Is Knowledge Management?

Knowledge management (KM) is the discipline of capturing, organizing, sharing, and applying what an organization knows so that the right knowledge reaches the right people - and, increasingly, the right systems - at the right time. It treats knowledge as an asset to be cultivated rather than something that happens to live in people's heads and scattered documents. Done well, it means a new hire can find how something is done, an expert's hard-won judgment does not walk out the door at retirement, and the same question is not solved from scratch in five departments.

Knowledge management is a mature field with decades of theory behind it, formalized to the point of having an international standard - ISO 30401:2018, Knowledge management systems - Requirements. What has changed is not the definition but the stakes: as organizations run on data and, now, on AI that must be grounded in trustworthy organizational knowledge, KM has shifted from a "nice to have" people-and-culture program to a foundation that determines whether analytics and AI can be trusted at all.

What Knowledge Management Is

At its core, KM covers a lifecycle of activities:

• Capture - getting knowledge out of individuals and processes and into a form others can use (documentation, recorded decisions, structured metadata).
• Organize - structuring it so it can be found: taxonomies, glossaries, ontologies, and relationships rather than an undifferentiated pile.
• Share - making it accessible across teams and time, with the right access controls.
• Apply - putting it to work in decisions, onboarding, problem-solving, and increasingly in AI systems.

ISO 30401 frames KM as a management system - like quality or information security - with leadership, culture, processes, and continual improvement, precisely because the recurring failure of KM is not technology but sustainability: knowledge bases that are created once and never maintained, and quietly rot into a liability.

Tacit and Explicit Knowledge

The most important distinction in KM is between two kinds of knowledge, because they demand entirely different handling.

• Tacit knowledge is experience, intuition, and judgment - the "knowing how" that is difficult to articulate. The seasoned operator who hears a bearing about to fail, the analyst who knows which data source to distrust. It is acquired through practice and hard to transfer except through direct contact.
• Explicit knowledge is knowledge that has been codified - the "knowing that" captured in documents, procedures, definitions, and databases. It is easy to store, search, and share, which is why most KM tooling targets it.

The tension is permanent: the most valuable knowledge is often tacit, and the most manageable is explicit. A great deal of KM is the ongoing effort to convert enough tacit knowledge into useful explicit form without pretending that everything important can be written down.

The SECI Model

The best-known framework for that conversion is the SECI model, developed by Ikujiro Nonaka in 1990 and refined with Hirotaka Takeuchi in their 1995 book The Knowledge-Creating Company. It describes knowledge creation as a continuous spiral through four modes:

• Socialization (tacit to tacit) - sharing experience directly, through mentoring, observation, and working side by side.
• Externalization (tacit to explicit) - articulating tacit knowledge into explicit form: concepts, models, and defined terms. This is the hardest and most valuable step, and it is exactly what a glossary or ontology captures.
• Combination (explicit to explicit) - connecting discrete pieces of explicit knowledge into new, more useful knowledge - the analytical heart of KM.
• Internalization (explicit to tacit) - absorbing explicit knowledge back into personal skill through learning and practice.

The model's enduring point is that knowledge is created by moving between tacit and explicit, repeatedly. Organizations that only ever document (externalize) without socializing or combining end up with dead archives; those that only socialize never scale their knowledge beyond the people in the room.

Knowledge Management in the Data and AI Era

Classic KM tooling - document repositories, wikis, intranets - solved storage but not structure. It made explicit knowledge available as prose, which humans can read but which does not connect: a definition in one document has no formal relationship to the data it describes, the process that uses it, or the person who owns it. That gap did not matter much when the only consumer was a human who could fill in context. It matters enormously now, for two reasons.

First, the volume of organizational knowledge has outgrown what anyone can navigate by browsing folders. Second, and more decisively, the newest consumer of organizational knowledge is not a person but an AI system - and an AI system cannot infer the unwritten context a human colleague would. To answer reliably, it needs knowledge that is structured: defined terms, explicit relationships, clear ownership, and provenance it can trace.

This is why modern knowledge management converges with data governance. The tools that make knowledge connected and machine-readable are the same ones that govern data: a business glossary that externalizes shared definitions, a catalog that inventories what exists and who owns it, and a knowledge graph or ontology that captures the relationships between concepts, data, and processes. Together they turn scattered explicit knowledge into a connected fabric - the kind of grounding that lets AI reason over an organization's knowledge instead of hallucinating around the gaps. It is worth being honest that this is an evolution, not a reinvention: the underlying disciplines - defining terms, cataloging assets, mapping relationships - are decades old. What is new is pointing them at machine consumers as well as human ones.

How Dawiso Fits

Dawiso is not a document management system or an intranet, and it does not try to be the place your policies and how-to guides live. Its role is the connected, governed layer of knowledge management - the part that makes organizational knowledge structured, findable, and trustworthy enough to ground both people and AI.

• Externalizing shared meaning. The business glossary is externalization in the SECI sense: it turns tacit, team-specific understanding of terms into explicit, agreed definitions everyone - and every system - can rely on.
• A catalog of what exists. The data catalog inventories data assets with owners and descriptions, so knowledge about the data estate is captured and maintained rather than held in a few people's heads.
• Connected, not just stored. By linking definitions, data, ownership, and relationships, Dawiso builds the kind of connected knowledge that a pile of documents cannot - and that AI needs to answer reliably.
• Sustained, not abandoned. Governance - ownership, stewardship, and change control - is what keeps this knowledge current, addressing KM's classic failure mode of the knowledge base that is built once and left to rot.

In short, Dawiso covers the explicit-and-connected end of knowledge management for the data domain, leaving documents and culture to the tools and practices built for them.

Conclusion

Knowledge management is the long-standing discipline of turning what an organization knows into an asset it can find, share, and apply - balancing hard-to-capture tacit knowledge against manageable explicit knowledge, and moving between them as the SECI model describes. Its principles have not changed, but its stakes have: as data and AI become the primary consumers of organizational knowledge, KM shifts from scattered documents toward connected, governed, machine-readable knowledge. A business glossary, a catalog, and a knowledge graph are how that shift happens - and they are the point where knowledge management and data governance become the same project.

Sources

• Wikipedia - SECI model of knowledge dimensions (Nonaka and Takeuchi; tacit and explicit knowledge).
• International Organization for Standardization - ISO 30401:2018, Knowledge management systems - Requirements.