Planning
Planning is the process of formulating and organizing intended actions or steps toward a goal, often accounting for dependencies, conditions, or constraints.
AI Agents in Action
Sequential planning is the organization of dependent actions into an ordered plan in which later actions follow earlier ones.
- How it is used
- Coordinates multistep tool use in assistants, autonomous agents, research tasks, and collaborative workflows.
- How deep it goes
- Explains sequential planning through prompts, JSON plans, local execution, context passing, evaluation, and feedback, while deferring detailed planner internals.
- Examples
- Uses Wikipedia retrieval and file saving, OpenAI Assistants, Nexus planners, and a time-travel reasoning problem.
- What you can do after
- Apply it via Nexus by building a prompt-based sequential planner that selects tools, executes subtasks, and incorporates feedback.
Build an AI Agent (From Scratch)
Planning is the construction of an ordered strategy or set of steps for completing a complex task.
- How it is used
- Directs agents through complex, multistep tasks by placing an ordered task list in their context before execution.
- How deep it goes
- Implements a one-plan-per-question Python tool with task statuses and full-list updates, while deferring multiple-plan and partial-update logic.
- Examples
- Traces a marathon-pace-to-Moon calculation through task creation, web search, and a checklist status update.
- What you can do after
- Build it from scratch as a create_tasks tool, then use task statuses to guide an agent’s next step.
Read how Build an AI Agent (From Scratch) explains planning →
AI Agents in Action, Second Edition
Planning is the process of constructing and organizing steps needed to achieve a goal.
- How it is used
- Coordinates an agent’s tool-using steps toward a multistep goal rather than treating each action as an isolated decision.
- How deep it goes
- Defines planning against decomposition and stepwise reasoning, then gives a checklist, storage approach, revision guidance, and bounded implementation example.
- Examples
- Uses a revised San Francisco itinerary and a Python planning agent with a checklist, structured answer, and turn limit.
- What you can do after
- Apply it via a sequential thinking MCP server to record and revise a plan while bounding execution and checking results.
Read how AI Agents in Action, Second Edition explains planning →
The Art of AI Product Development
Online planning is the dynamic generation and revision of an action plan while an agent is executing a task.
- How it is used
- Uses language-model agents to coordinate multi-step tasks involving tools, decisions, feedback, and changing execution conditions.
- How deep it goes
- Explains decomposition, reflection, and explicit flows, but stops short of detailed implementation or reliable long-horizon replanning.
- Examples
- Demonstrates roadmap updates through retrieval, web research, user-feedback analysis, strategic alignment, and subtask reflection.
- What you can do after
- Evaluate and design agent workflows by balancing adaptive planning against explicit controls for errors, latency, and user disruption.
Read how The Art of AI Product Development explains planning →
Investing for Programmers
A planning pattern is an agent architecture in which an agent dynamically constructs and executes a sequence of actions toward a complex goal.
- How it is used
- Structures AI agents that pursue complex goals through dynamically assembled action sequences.
- How deep it goes
- Defines planning within a pattern taxonomy but does not detail planning algorithms, representations, or execution mechanics.
- Examples
- Uses a reference workflow for an AI agent as the main visual artifact.
- What you can do after
- Evaluate/understand planning as a way to organize agent behavior around ordered actions and complex goals.
Spring AI in Action
Agent planning is the capability of selecting a path through available actions to achieve a desired goal.
- How it is used
- Uses dynamic action selection to achieve goals without a predetermined workflow.
- How deep it goes
- Demonstrates planning through execution traces rather than formal planning algorithms.
- Examples
- Uses Azul questions about player count and game mechanics to show different action paths.
- What you can do after
- Evaluate an agent’s autonomy by observing which actions it selects for a goal.
Related concepts
- chain-of-thought promptingused by · Chain-of-thought prompting is a technique that elicits or guides a model to use intermediate reasoning steps before reaching an answer.
- 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.
- project life cyclepart of · A project life cycle is the sequence of phases through which a project progresses from initiation through completion and delivery.
- decompositioncontrasts with · Decomposition is the process of dividing a complex whole into smaller constituent parts that can be handled or analyzed separately.
- loopcontrasts with · A loop is a programming control-flow construct that repeatedly executes a block of instructions according to an iteration rule or continuation condition.
- flow engineeringalternative to · Flow engineering is the deliberate design of an explicit sequence of operations for an AI workflow rather than relying entirely on autonomous planning.
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.