A Brighter Future for KM: Building an AI-Enhanced Knowledge-Sharing Capability

Knowledge is fluid, contextual and everywhere. How can we create an organisational capability to map, connect and synthesise enterprise knowledge in changing times?

Knowledge management has long been burdened by a paradox. The promise of unlocking organisational wisdom — to connect the right information, to the right people, at the right time — is as compelling as it is elusive.

Despite decades of investment in repositories, taxonomies, and intranet systems, KM initiatives often faltered, falling victim to outdated technology, cultural resistance, or sheer complexity. The result? Vast reserves of untapped institutional knowledge languish in silos, disconnected from the decisions and innovations they might inform.

Why did this happen? Traditional KM approaches relied heavily on human effort: manually tagging documents, maintaining databases, and ensuring compliance with ever-evolving organisational structures. These systems were static, brittle, and ill-suited to the dynamic realities of modern enterprises. Worse still, they lacked the ability to adapt and learn, making them unresponsive to changing organisational needs and user behaviours.

But could AI usher in a new era of enterprise knowledge management?

With its ability to process vast quantities of unstructured data, identify patterns, and adapt in real time, AI offers a fresh approach to knowledge management. By automating labour-intensive tasks like content tagging and relevance matching, AI removes the bottlenecks that stymied earlier KM efforts. Natural Language Processing (NLP) allows systems to understand context and intent, turning a simple query into a gateway for actionable insights. Knowledge graphs map relationships between people, resources, and concepts, creating dynamic connections that evolve alongside the organisation.

But AI’s promise in KM is not just about efficiency - it’s about transformation. AI doesn’t merely catalogue knowledge; it elevates it. By embedding predictive analytics, collaborative tools, and personalised recommendations, it redefines how knowledge is shared, accessed, and applied. What was once a static repository becomes an intelligent, adaptive system — a true enabler of innovation and decision-making.

Potential application areas for AI-enhanced KM

There are so many potential application areas for AI-enhanced KM that are both elegant and useful for modern enterprises, and which could play a strategic role in times of change.

For example:

• Enterprise LLMs. It is becoming easier and more affordable to train LLMs using your own knowledge stock and flows, which creates all kinds of possibilities for knowledge agents and other applications. This is likely to become a foundational capability in the near future that will reduce the risks of data leakage or hallucinations inherent in using generic public LLMs for specific enterprise purposes.

• Personal knowledge agents. A simple agent that gradually learns your interests, areas of work and content preferences could be very helpful in keeping us in-the-know on both external developments and new information or knowledge within the workplace that we might not be aware of. We will give them research tasks and use them to quickly get an overview of new or emerging areas of knowledge. They might even help map and organised our own content. Pre-Google, when search engines were just emerging, very simple search agents were available on the web, but they were quite limited; now we can finally do them properly.

• Knowledge synthesis and mapping. With knowledge scattered across systems and repositories, manual knowledge audits or mapping projects have always been sysphean tasks, and of course knowledge is not fixed, but gets constantly re-interpreted and re-combined over time. The ability to synthesise knowledge has already been demonstrated very well by leading LLMs, and this will be even more useful in an enterprise context. But the ability to rapidly map and re-map a knowledge domain or knowledge graph from different perspectives is potentially a very valuable capability in times of rapid change. From ‘what do we know about X?’ use cases to re-mapping existing knowledge based on new developments, there are plenty of use cases for this knowledge super-power - even simple ones such as employee onboarding (”guide me through what a new person might need to know about this company’s work”) and even project start-up support (”if I am trying to build X, who should I talk to and what should I read before I begin?”).

• Emerging insights discovery. With locally-trained LLMs and the ability to synthesise existing knowledge, it becomes easier to identify connections, correlations or ‘echoes’ between previously separate knowledge domains. This can assist with finding relevant or adjacent emerging insights that might be useful to an organisation. Taking the example of European automotive and energy firms, in both cases some leading companies failed to see the importance of changes happening in plain sight, but adjacent to their markets; and they paid a heavy cost for this myopia. Pro-active, perhaps agent-based emerging insights discovery could be useful in such cases and alert analysts and leaders to new possibilities or risks.

• Decision support and dialogic agents. There has been a lot of talk of data-based decision-making, but in most difficult decisions, there are many factors to consider in addition to raw data. For leaders at every level, the ability to query data and knowledge stores could be useful, but even more so in a dialogic context - i.e. having a conversation to explore different angles or potential outcomes. This is something that even today’s LLMs are actually rather good at, acting as a kind of coach, helping people formulate and explore their own thinking in advance of taking action. Just as we will probably have personal knowledge agents working away to keep up informed, we will probably also have decision support coaching agents to help interrogate our thinking and ideas in much the same way we already do with our people and teams.

So, let’s explore the underlying capability of AI-enhanced enterprise knowledge management to understand what is needed to support the kinds of outcomes we think are possible. From breaking down silos to fostering collaboration and surfacing insights that were once invisible, AI is not just reviving KM’s original promise — it is also redefining it for the future of work.

Identifying Key Components

Every digital business capability can be broken down into the ingredients and components needed - like a recipe - to make it more actionable and help answer the following key questions:

• What components does the organisation already have in place, and which are missing?

• What related new technologies or initiatives are already planned?

• What services - and therefore skills - need an upgrade or re-training?

Let’s dive deeper to discover how to evolve this transformative capability.