Five Lessons from Twenty-Five Years of Designing Knowledge Solutions

Looking back across scientific publishing, competitive intelligence, enterprise knowledge platforms, and AI-enabled knowledge systems, one observation stands out:

The fundamental challenges have remained remarkably consistent.

Technologies have changed dramatically.

User expectations have evolved.

Artificial intelligence is reshaping the landscape.

But the underlying questions are often still the same:

• How do we help experts navigate complexity?
• How do we transform information into actionable knowledge?
• How do we design solutions that are well adopted and genuinely support better decisions?

The following lessons have consistently proven important.

Organizations often become fascinated by new technologies: Search, semantic technologies, knowledge graphs, generative AI. All of these capabilities can create value. But technology should never be the starting point.

The starting point must always be a concrete business problem and a clear understanding of user needs. Successful knowledge solutions are problem-driven rather than technology-driven.

Users are rarely homogeneous. Different personas have different objectives, time pressures, evidence requirements, and ways of working. Designing effective knowledge solutions therefore requires understanding workflows rather than simply collecting requirements.

A solution that works extremely well for one user group may be completely inadequate for another. Context matters.

This lesson becomes increasingly relevant in the age of AI. Users do not only need answers. They need confidence that these answers are reliable.

Trust requires:

• Transparency
• Traceability
• Explainability

Especially in scientific and innovation environments, trust remains the foundation of adoption.

No matter how sophisticated a platform may be, value is only created when people use it. Communication, training, communities, ambassadors, and management support are therefore not optional activities. They are integral parts of every successful knowledge initiative.

Adoption should be designed with the same care as technology.

Information alone rarely creates value. Value emerges when information helps people make better decisions. This may involve choosing a research direction, assessing intellectual property risks, identifying emerging trends, or evaluating strategic opportunities.

Knowledge solutions therefore should not primarily be measured by the number of documents indexed or the sophistication of their algorithms. Their true value lies in whether they help experts make better decisions under uncertainty.

The future of knowledge management will undoubtedly be shaped by artificial intelligence. However, the organizations that will benefit most from these developments are unlikely to be those with the most advanced technologies alone.

They will be the organizations that successfully combine:

• Trusted content
• User-centered design
• Effective governance
• AI capabilities that fit real workflows.

Technology will continue to evolve. The principles of effective knowledge solutions are likely to remain surprisingly constant.

Helping experts transform information into understanding and understanding into action has been a consistent challenge for more than twenty-five years. I suspect it will remain so for many years to come.