Parallel flow
A parallel flow is a workflow pattern in which independent activities or jobs execute concurrently rather than sequentially.
AI Agents in Action, Second Edition
A parallel flow is a multi-agent pattern in which independent agents execute subtasks concurrently and a downstream step combines their outputs.
- How it is used
- Coordinates independent agents on concurrent paths before a downstream step combines their outputs.
- How deep it goes
- Defines the pattern through a two-middle-agent diagram and selection table, without specifying implementation or failure handling.
- Examples
- Uses a diagram with two middle agents executing simultaneously and a table identifying when parallel delegation fits.
- What you can do after
- Evaluate/understand it by choosing parallel delegation when multiple tasks or agents can run together to improve performance.
Read how AI Agents in Action, Second Edition explains parallel flow →
Build an AI Agent (From Scratch)
A parallel workflow runs multiple constituent agents or steps concurrently and combines their results.
- How it is used
- Runs independent analyses concurrently when combining diverse perspectives is more useful than sequential processing.
- How deep it goes
- Explains asynchronous execution, context boundaries, duplicate-event removal, result merging, and workflow composition.
- Examples
- Uses optimistic, pessimistic, and realistic startup analysis plus security, performance, readability, and multi-subtopic research scenarios.
- What you can do after
- Build it from scratch as an asynchronous Agent-compatible workflow that runs agents concurrently and merges their results.
Read how Build an AI Agent (From Scratch) explains parallel flow →
Related concepts
- pipelinealternative to · A pipeline is an ordered or coordinated sequence of connected processing stages or tasks that passes data, work, or results toward an outcome.
- AI agentkind of · An AI agent is a computational system that uses artificial intelligence to interpret information, select actions, and pursue goals through interaction with an environment.
- matrix strategyproduced by · A matrix strategy expands a workflow job into multiple executions covering combinations of configured parameter values.
- asyncio.gatherimplemented by · asyncio.gather is a Python asyncio function that runs multiple awaitable objects concurrently, waits for them to complete, and collects their results.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.