Purpose: The growing prominence of knowledge workers in contemporary organisations has led to a considerable amount of research into their role and activities, however, despite this growing interest, there remains a lack of clarity regarding the relationship of knowledge workers to management. This paper aims to respond by investigating the features of knowledge work that impose barriers to traditional managerial control methods and exploring the extent to which alternative approaches to influence are available to knowledge work managers, and the circumstances under which these different approaches are indicated. Design/methodology/approach: The paper utilised agency theory to investigate the implications of knowledge asymmetry in principal-knowledge worker exchanges, and argues against the utility of models of control based on principles of scientific management on which agency models are typically based, towards models of control based on reciprocity and commitment. Findings: The paper has two main outputs. First, it argues that situations in which knowledge-based specialists are hired to perform tasks requiring expertise can be viewed as principal-agent exchanges (Mills) and are therefore open to analysis through an agency theory lens. The second endeavour is to discuss the implications of knowledge asymmetry in principal-knowledge worker relationships for managerial control and influence, and posit alternative modes of managerial control based on normative influence and valiant rewards. Further, it develops propositions relating to the factors likely to moderate the utility of each managerial influence strategy. Originality/value: By undertaking this investigation, the paper responds to calls by Frenkel et al. to understand process and dynamics of control in managing knowledge workers. It moves away from models of control based on principles of scientific management on which agency models are typically based, towards models of control based on reciprocity and commitment. It develops testable propositions regarding specific sources of influence in knowledge work and the circumstances under which the employment of these influence strategies is indicated. This responds to calls by authors such as Tampoe; and Lord to develop a detailed understanding of the manner and extent to which influence strategies need to complement specific knowledge worker characteristics. (Contains 1 figure.)