liveBook · Manning

Iterative debate pattern

As explained in AI Agents in Action, Second Edition

An iterative debate pattern is a multi-agent collaboration in which agents repeatedly propose, critique, and revise solutions toward a vetted result.

The iterative debate pattern has two or more peer agents propose, challenge, and refine responses toward consensus. Use it when a response deserves a second or third check, not when the task is trivial or simple execution will do. section 4.1.3

Do not confuse debate with iterative critique. In critique, one critic has authority over one worker’s output and sends it back for revision until it meets the goal. In debate, peers argue toward consensus; no single agent controls the result. section 4.1.3

Table 4.4 Architectural patterns overview

Pattern

When to use

When not to use

Sequential pipeline, agent to agent

Well-defined, ordered tasks need to be executed in sequence.

Not all agents need to be executed every time.

Parallel delegation

Multiple tasks or agents can be executed in parallel to improve performance.

Hierarchical coordination

The user interface agent delegates to multiple agents on behalf of the user.

No need for complex decision-­making or command of agents

Iterative debate and refinement

You want to review and double- or triple-check responses.

The problem is trivial or aligns well with simple execution.

Voting/best-of-N (ensemble)

Ambiguous problems that may have multiple interpretations or facets

Unambiguous problems

Role-playing collaboration

High creativity and novel outputs are desired and align well with the principal goals.

Real-time systems or cases with low desired creativity

Conditional routing (branching)

Customer support systems that may require human approvals or intervention

Peer-to-peer network

Decentralized command increases system robustness and reduces or eliminates complete failure.

Compare iterative debate and refinement with other coordination patterns by checking when each is useful—and when its extra coordination is unnecessary. section 4.1.3

Start with the simplest coordination pattern that works. Add debate when repeated review has a clear benefit, and set tighter constraints when the stakes are higher, such as customer-facing output or irreversible actions. section 4.1.3

How AI Agents in Action, Second Edition covers it

How deep it goes
Defines debate by peer relationships and consensus, contrasts it with one-to-one critique, and gives selection guidance without an implementation walkthrough.
Examples
Uses a single critic revising a worker’s output as a contrast and a pattern-selection table as a decision aid.
What you can do after
Evaluate/understand it as an option for repeated peer review, and reject it when simple execution is sufficient.
How it connects
Contrasts iterative critique and relates debate to sequential flow, parallel delegation, hierarchical coordination, and voting/best-of-N.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of AI Agents in Action, Second Edition. The section links point to the parts of the book it draws on.