The question of who manages AI agents at work is landing on HR’s desk, and most HR functions aren’t ready for it. Accenture’s UK CEO has stated directly that organizations will need to onboard and govern AI agents as a class of workforce member — not as software tools, but as entities with defined activation conditions, performance expectations, and decommissioning procedures. The shift puts HR in unfamiliar territory: building lifecycle processes for workers that don’t get paid and don’t sit at desks, but carry real training, governance, and oversight costs.
The data behind the prediction: 84% of executives surveyed by Accenture expect to have AI agents in their organizations within three years. Only 26% of employees are currently receiving any training on how to collaborate with AI systems. That gap — between deployment timelines and workforce readiness — is the HR problem. If agents arrive before governance frameworks do, the costs show up later as audit findings, remediation projects, or talent friction from staff who weren’t prepared to work alongside autonomous systems.
Some organizations are already moving: Accenture reports clients including AI agent capacity in formal workforce plans, treating them as a headcount category that requires resourcing decisions even if it doesn’t require payroll. CHROs already navigating the AI confidence gap between leadership and frontline employees will recognize the dynamic — the governance deficit compounds when agents are deployed into a workforce that hasn’t been prepared to receive them.
The practical implication: HR functions that define AI agent lifecycle now — what activation looks like, how performance is monitored, when an agent is retired — will be better positioned when regulators and boards start asking those questions. The organizations waiting for IT to sort it out are likely to find that the HR costs land on them anyway.
Source: Accenture Banking Blog — Agentic AI Impact on Workforce Change
