Your next workforce might include people who never appear on the org chart.
I think we need to change the way we think about AI governance. Think of agents as a digital workforce. An agent should have an identity. It should have a role and job description. It should have defined access rights. It should have a reporting line and an accountable human owner.
And perhaps we will even have the equivalent of career levels: Junior agent: narrow permissions, close supervision, limited autonomy. Senior agent: broader scope, more tools, greater autonomy, higher potential impact. A promotion would then mean granting additional capabilities or access only after the agent has demonstrated sufficient reliability and passed the appropriate controls.
The analogy becomes particularly interesting when agents start managing or delegating tasks to other agents. Who is accountable for the hierarchy? Can one agent grant permissions to another? How far can delegation extend? When must a human intervene? And what happens when an agent changes role?
This is where established AI governance frameworks can give us a useful foundation. The NIST AI Risk Management Framework, for example, structures AI risk management around four functions: GOVERN, MAP, MEASURE and MANAGE. Its accompanying NIST AI RMF Playbook translates these into suggested actions covering areas including governance and oversight, human oversight, monitoring, privacy, security, drift and decommissioning.
For organisations deploying agents, I think AI governance will increasingly start to look like a combination of: AI Governance + Identity amp; Access Management + Organisational Design Maybe the future AI register will look less like an inventory of software and more like an org chart for a digital workforce.