Yes - knowledge graphs are useful

They shine when a pharma organization wants to uncover hidden relationships between molecules and side effects, or when an intelligence agency needs to connect dots across complex data networks .

These are powerful, high-stakes use cases where explainability, traceability, and reasoning truly matter.

However…

For most organizations, investing in a full-scale knowledge graph may not be the smartest or most economical move.

In 80–90% of cases, lighter approaches - good metadata, clear taxonomies, and AI-based semantic search or RAG assistants  - already provide the answers we seek.

They connect silos, find relationships, and support decision making without the heavy infrastructure or ongoing maintenance that graphs require.

Knowledge graphs are not irrelevant - they’re simply one instrument in a much larger KM-AI orchestra.

The real wisdom lies in choosing the right level of structure for your context, cost, and purpose.

So before starting a “graph project,” it might be worth asking:

Do we need deep, accurate reasoning… or just better connection?