Platform
Intelligent operating models: people and agents at work
The hard question is not what agents can do. It is how the org changes when they do it — and who owns what.
Deploying an agent is a technical act. Deciding how people, agents, and governance share the work is an organisational one — and it is the part most programmes leave until it is too late to do well.
Design the handoffs first
The value of an autonomous system lives in its handoffs: which actions an agent takes alone, which need a person, and how control passes between them. Design those deliberately and the system is legible. Leave them implicit and you get an automation nobody trusts and everybody second-guesses.
People move up the value chain
Done right, autonomy does not remove people — it moves them. Off repetitive triage, onto the exceptions and the judgement calls. The operating model has to make that shift explicit, or the org fights the system instead of using it.
Automate the routine, and you don't shrink the team — you promote it.
Governance is part of the product
Who can change an agent's guardrails? Who reviews the audit trail, and how often? An intelligent operating model answers these before go-live, so oversight is a designed feature rather than a reaction to the first incident.