A human-agent operating model begins with the work and its decisions rather than assigning agents to yesterday's job descriptions. This matters now because technology rollout can define informal roles and escalation paths before the organization has consciously designed them.
Deloitte describes agent and multi-agent adoption as an early-stage management problem with unresolved questions about workforce design. The article asks leaders to reconsider jobs, decision rights, workflows, and the relationship between human and digital workers. The mechanism is task decomposition combined with explicit authority: routine lanes can be delegated while ambiguous, exceptional, or high-consequence decisions retain human ownership.
Leaders should preserve learning paths, name accountable owners, and measure how supervision load changes as agent autonomy expands. The evidence is Deloitte's forward-looking organizational analysis, not a mature empirical standard for every occupation or industry. Watch whether firms publish stable decision-rights models and career pathways for employees who supervise increasingly capable agents.