Leadership is often associated with judgement, authority, vision and decision-making, but in a knowledge-intensive institution there is another capability that matters just as much and is discussed far less often: knowing where the relevant knowledge lives. The larger and more specialised an organisation becomes, the less plausible it is that any one leader can personally understand every domain in sufficient depth. Public institutions deal simultaneously with legal interpretation, operational delivery, finance, technology, public behaviour, political constraints, risk, data, professional standards and local context. The cognitive challenge of leadership is therefore not simply to know more. It is to navigate an organisation in which what needs to be known is distributed across many people, units, systems and forms of expertise.

This changes what competence at the top of an institution can reasonably mean. A leader does not need to possess every answer, but they do need some understanding of where answers are likely to be found, which forms of expertise matter for a particular problem and when the apparent answer offered by one part of the organisation is incomplete because another relevant perspective has not yet entered the conversation. Leadership knowledge is therefore partly second-order knowledge: not only knowledge about the substantive issue, but knowledge about the institution’s own knowledge landscape.

This distinction matters because institutional knowledge is rarely organised in a way that mirrors the problems institutions face. Organisational charts divide work into units, professional systems divide expertise into disciplines and administrative processes separate responsibilities according to formal mandates. Real problems often cut across all of these boundaries. A failure in a public service may involve technology, procurement, frontline practice, regulation and user behaviour at the same time. A policy decision may require legal expertise, implementation knowledge, economic analysis and understanding of local conditions. No single organisational unit necessarily contains the whole picture, so leadership becomes partly a problem of finding and assembling the right fragments of institutional knowledge.

Knowing who knows what should not be confused with maintaining a large personal network. Informal relationships can certainly help leaders locate expertise quickly, but personal familiarity is an unreliable substitute for institutional epistemic navigation. Networks are often uneven. Some people are visible because they are close to senior decision-makers, while others hold highly relevant knowledge from positions that receive little organisational attention. A leader who relies primarily on personal contacts may repeatedly consult the same trusted voices and mistake familiarity for coverage. The deeper capability is not simply knowing many people; it is understanding how expertise is distributed and having reliable ways to reach it beyond one’s immediate social field.

The problem becomes particularly important in specialised institutions because specialisation creates both capability and distance. As expertise deepens, professional communities develop their own language, assumptions, evidence standards and ways of framing problems. This improves local judgement but makes it harder for others to know what expertise exists, when it should be consulted and how it should be interpreted. A senior decision-maker may therefore face a paradox: the institution contains more knowledge than ever, while the cognitive task of locating and connecting that knowledge becomes increasingly difficult.

In this environment, effective leadership includes the ability to ask not only, “What do I think?” but also, “Who in this institution has encountered this before?”, “Which unit sees a part of this problem that we do not?”, “Whose expertise is missing from this discussion?” and “Where would a competing interpretation come from?” These questions do not weaken authority. They improve the conditions under which authority is exercised. A leader who understands the limits of personal knowledge can use the institution’s distributed expertise without pretending that every decision can be delegated to specialists.

This last point is important because epistemic navigation is not the same as surrendering judgement. Experts know domains, but leadership often involves decisions that span domains and require trade-offs among competing forms of knowledge. A legal interpretation may be correct but operationally difficult. An analytically optimal solution may create political or ethical problems. A technically feasible intervention may be institutionally illegitimate. Knowing who knows what helps the leader bring the relevant perspectives into the decision, but it does not eliminate the need to judge among them. Expert location is therefore a precondition for informed judgement, not a replacement for judgement.

The quality of this navigation also depends on whether the organisation makes expertise discoverable. In many institutions, knowledge about who knows what is itself highly informal. Long-serving staff know whom to call, which team has seen a similar issue before and which individual understands how a supposedly simple process actually works. New leaders or employees may have access to the same organisational chart but not to this hidden map. When those informal knowledge maps disappear through turnover, the institution can retain all of its formal expertise while losing the ability to find it efficiently.

For this reason, epistemic navigation is not only an individual leadership skill. It can be supported institutionally. Clearer responsibility maps, communities of practice, searchable expertise directories, cross-functional forums, boundary-spanning roles and deliberate mechanisms for escalating unusual problems can all make distributed knowledge easier to access. The purpose of such mechanisms is not to centralise all knowledge but to make decentralised expertise more reachable when the organisation needs to combine it.

Technology can contribute to this capability, especially as institutions become more complex. Digital systems can help identify subject-matter expertise, connect related cases, route questions to relevant units or reveal where similar problems have appeared elsewhere. Artificial intelligence may eventually make it easier to search across institutional knowledge and identify expertise that is not obvious from formal structures. Yet technological discoverability creates its own risks. Expertise that is well documented may become easier to find than tacit or experiential knowledge; formal job descriptions may overstate what someone actually knows; and systems may reproduce existing organisational hierarchies by making already-visible experts even more visible. Technology can assist epistemic navigation, but it does not remove the need for judgement about what kind of knowledge is actually required.

Leadership becomes especially difficult when nobody clearly owns the relevant knowledge. Emerging problems often appear between established categories. An issue may be partly technical, partly behavioural and partly regulatory, yet fit comfortably within none of them. In those cases, knowing who knows what also means recognising when no existing expert or unit holds an adequate answer. The leadership task then changes from locating knowledge to constructing a temporary knowledge configuration: bringing partial perspectives together, commissioning new analysis, seeking external expertise or creating a space in which the institution can learn what it does not yet know.

This is why epistemic navigation becomes more valuable as institutional complexity grows. In a small organisation, direct familiarity may be sufficient. In a large institution, knowledge topology becomes a design problem. What matters is not only how much expertise exists but whether decision-makers can reach the relevant expertise at the right moment, whether peripheral knowledge can enter central judgement and whether multiple perspectives can be combined before a decision becomes irreversible.

The strongest leaders in such environments are not necessarily those who appear to know the most. They are often those who understand how little any one person can know and who are able to activate the wider cognitive resources of the institution without confusing access to expertise with abdication of responsibility. They know that a good question directed to the right part of the organisation may be more valuable than a confident answer generated from an incomplete view.

Leadership in a distributed knowledge system is therefore partly the art of knowing who knows what. It is the ability to navigate an institutional landscape in which expertise is scattered, specialised and unevenly visible, and to connect that expertise to the problems for which it matters. The leader does not need to become the institution’s repository of knowledge. The deeper capability is to understand how institutional knowledge is distributed well enough to know where to look, whom to involve and when the organisation’s existing knowledge map is no longer sufficient.