What is multi-agent system?
A multi-agent system is a set of specialised AI agents that work together on a task, each handling the part it is set up for, such as billing, bookings or checking another agent’s work.
Why it matters
One agent given every instruction, every tool and every policy tends to become harder to control and test. Splitting the work lets each agent have a narrow job, a short set of instructions and only the tools it needs. That can make behaviour more predictable. The trade-off is added complexity, since the agents must pass information to each other correctly.
How it works
A routing agent usually reads the request and decides which specialist should handle it. The specialist does its part and returns the result or passes the conversation on with a summary. Some systems add a reviewing agent that checks a reply against policy before it is sent. The customer normally experiences one continuous conversation, not a series of separate agents.
Example
A customer messages a telecoms provider to query a charge and then asks to upgrade a handset. A routing agent sends the first part to a billing agent, which explains the charge from the invoice. The conversation then moves to a sales agent with a short summary attached. A reviewing agent checks the upgrade offer against current terms before the customer sees it.
