??PeopleAI agentstraining: reward shapesbehaviour before workincentivesnormsrolesa person with an agentharnessroles · permissionsmemory · checksagent to agentan agent on its ownnormsincentivesPeopleAI agentsa person with an agentharnessagent to agentan agent on its own

Agentic Resource Management

Organisations have always managed behaviour, not just output. Now part of the workforce is artificial. Agentic Resource Management studies which levers shape the behaviour of people and AI agents, and how both are held to account: a person with an agent, an agent on its own, and agent to agent.

  • Person
  • AI agent
  • Delegation
  • Harness
  • Lever
  • Open question
The thesis

Managing work has always meant managing behaviour

No organisation can write down everything a worker should do. Where instructions run out, managers rely on incentives, norms, culture and judgement to shape what people do. AI agents now take on work with the same gap between instruction and action: they plan, act and make choices no one fully specified. Agentic Resource Management asks which of the ways we shape human behaviour carry over to agents, which fail, and what has to be invented.

A person with an agent

People giving agents goals, shaping how they work, and judging and raising the quality of what they produce.

An agent on its own

An agent working through a task by its own choices, within the permissions and limits its harness sets.

Agent to agent

Agents assigning work to one another, checking each other and passing results along a chain, where norms and accountability must hold without a person at every step.

In short: ARM is the study and practice of directing, shaping and governing the behaviour of people and AI agents entrusted with goals and discretion, so that their work stays within intent and someone answers for it. Read the full definition