Chapter 5: Prompts in action
This introduction previews a project-management demonstration that applies prompt engineering first and problem formulation second, while explaining why a traditional cross-platform prompting approach is used.
After this chapter you will be able to
- Chapter demonstration approach. Describe the chapter's two-part approach to creating practical, work-ready project-management deliverables with AI: basic prompt engineering first, followed by problem formulation to streamline the work.
- Cross-platform prompting foundation. Explain why the demonstration uses a traditional prompting approach that works across chatbot platforms: it can be applied regardless of model and builds a deeper foundation in effective prompt-engineering principles.
- Project management capability gap. Explain why IT professionals may need to manage projects themselves and why informal tools alone may not adequately organize a project. State that formal project-management techniques offer benefits but may be unfamiliar without PM training.
- Zero-shot prompt limitations. Explain why the initial question, “How can I be a good project manager?”, is a zero-shot prompt and why its broad wording produces wide-ranging but insufficiently useful output for the immediate task.
- Contextual prompt refinement. Describe how the revised prompt improves the output by stating the user's role, the small IT-project context, the absence of budget responsibility, and the request for tasks to complete.
- Recursive prompt context. Explain how a recursive prompt builds on information and outputs already established earlier in the same chat conversation instead of restating all details.
- Documentation follow-up example. Describe how the example uses the follow-up request for supporting documents to turn an already established small-project context into targeted project-management documentation suggestions.
- Context-aware template generation. Describe how the chatbot generates a simplified project-plan template from the user's request while retaining earlier conversational context, including suggested document types and exclusions.
- Template refinement and reuse. Explain that the generated template should be reviewed for missing or unsuitable sections, modified as needed, then saved as reusable actionable content for future projects before asking the chatbot to fill it with project details.
- Conversational template completion. Describe how to provide a chatbot with project information in natural language and ask it to fill an existing project-plan template. Explain that the source information can be a write-up, bullet points, or spreadsheet rows, although this example supplies it conversationally.
- Migration plan derivation. Trace how the Oracle Database Migration description becomes a project plan with an objective, milestones, sequential dependencies, resource assignments, communications, and risks. State that the migration moves Oracle 19c to 23ai within a two-hour downtime window.
- Draft refinement limits. Explain that AI can rapidly generate project-management suggestions, required documents, and a draft template output, but that the result still needs review and tweaks. Distinguish the convenience of natural-language prompting from its weakness when the plan requires many detailed, structured fields such as risk classifications and task dates.
- Top-down problem formulation. Distinguish the top-down problem-formulation approach from the bottom-up approach used in section 5.1. Explain that problem formulation starts with a known desired outcome and planned steps, so an experienced practitioner guides the AI with a map rather than asking it to discover the path.
- Structured project template prompt. Describe how the example prompt creates project documentation from spreadsheet data. State that it assigns the AI an expert senior-project-manager role, specifies the required template sections, and instructs it to use dependency and milestone flags to populate their own sections rather than placing those flags in the task list.
- Expert-guided AI productivity. Explain why this is problem formulation rather than prompt engineering alone: the user tells the AI what to create instead of asking what to do or trying to learn from it. Describe how detailed structured input and preplanning reduce prompting effort, lower error risk, and make the AI a productivity tool whose output aligns with specific objectives.
What it builds on
- Prompt engineering (chapter 4) · The idea of prompt engineering as a way to direct an AI model.
- Problem formulation (chapter 4) · The concept of problem formulation.
- Summary comparison of AI assistants (chapter 2) · The existence of multiple AI assistants with differing capabilities.
- The mechanics of a good prompt (chapter 4) · Principles for constructing a good prompt.
- Free vs. premium accounts (chapter 3) · The distinction between free and premium chatbot access.
- Zero-shot prompting (chapter 4) · The definition and use of zero-shot prompting.
- Key principles in prompt construction (chapter 4) · The principles used to construct effective prompts.
- Explicit constraints (chapter 4) · Using explicit constraints to guide a model's response.
Section by section
Introduction
This introduction previews a project-management demonstration that applies prompt engineering first and problem formulation second, while explaining why a traditional cross-platform prompting approach is used.
- Chapter demonstration approach: Describe the chapter's two-part approach to creating practical, work-ready project-management deliverables with AI: basic prompt engineering first, followed by problem formulation to streamline the work.
Why it matters: This roadmap helps you recognize how the later examples move from an initial prompt toward a more efficient solution. - Cross-platform prompting foundation: Explain why the demonstration uses a traditional prompting approach that works across chatbot platforms: it can be applied regardless of model and builds a deeper foundation in effective prompt-engineering principles.
Why it matters: A transferable prompting foundation lets you use the method even when the chatbot or model changes. - OpenAI model trade-offs: Distinguish the model characteristics described for ChatGPT-4: 4o emphasizes efficient real-time performance, 4o mini trades some capability for speed and lower API cost, and paid o1 and o3 spend longer analyzing and planning requests for consistently superior results.
Why it matters: Knowing these trade-offs helps you compare the same exercise across models and interpret differences in speed, cost, and output quality.
Key terms: prompt engineering, problem formulation, practical, work-ready deliverables, traditional approach, chatbot platform, effective prompt engineering principles, ChatGPT-4, 4o, 4o mini, o1
5.1 Prompt engineering in action
This introduction frames project management as a common IT responsibility where AI can help practitioners apply formal practices despite limited project-management training.
- Project management capability gap: Explain why IT professionals may need to manage projects themselves and why informal tools alone may not adequately organize a project. State that formal project-management techniques offer benefits but may be unfamiliar without PM training.
Why it matters: Recognizing this capability gap explains why structured prompting can help you organize project work more effectively.
Key terms: project management, formal project management techniques, Excel, Gantt chart, PM training
5.1.1 Basic prompting
This section demonstrates how adding concise, relevant context and constraints to a basic zero-shot prompt turns broad AI advice into actionable guidance for a small IT project.
- Zero-shot prompt limitations: Explain why the initial question, “How can I be a good project manager?”, is a zero-shot prompt and why its broad wording produces wide-ranging but insufficiently useful output for the immediate task.
Why it matters: Recognizing when a prompt is too open helps you revise it before spending time sorting through unfocused AI output. - Contextual prompt refinement: Describe how the revised prompt improves the output by stating the user's role, the small IT-project context, the absence of budget responsibility, and the request for tasks to complete.
Why it matters: A small amount of situational detail can make AI output substantially more applicable without making the prompt wordy. - Four prompt quality principles: Distinguish clarity, conciseness, relevance, and specificity as applied to the revised prompt, including why concise prompts can still contain enough context to target a useful response.
Why it matters: These four principles give you a practical way to assess and improve a prompt before relying on its output.
Key terms: zero-shot prompt, problem formulation, relevance, specificity, clarity, conciseness, constraints
5.1.2 Recursive prompting
This section shows how recursive prompts use prior chat context to request relevant follow-up work without repeating earlier details.
- Recursive prompt context: Explain how a recursive prompt builds on information and outputs already established earlier in the same chat conversation instead of restating all details.
Why it matters: Using the ongoing context can make follow-up requests more natural, focused, and relevant. - Context boundaries and focus: Describe why recursive prompting works best when the discussion stays focused and relevant, and state that beginning a new chat wipes out the conversation context.
Why it matters: Knowing the context boundaries helps you avoid relying on information the chatbot no longer has or mixing unrelated work into one conversation. - Documentation follow-up example: Describe how the example uses the follow-up request for supporting documents to turn an already established small-project context into targeted project-management documentation suggestions.
Why it matters: This example demonstrates how a short context-dependent request can produce actionable, task-specific output.
Key terms: recursive prompts, context, chat session, previous prompts, focused discussions, relevant context, new chats, supporting documents, Project Charter, Project Plan
5.1.3 Template creation
This section shows how to ask a chatbot for a simplified, reusable project-plan template, refine or save its output, and prepare it to be filled with project details.
- Context-aware template generation: Describe how the chatbot generates a simplified project-plan template from the user's request while retaining earlier conversational context, including suggested document types and exclusions.
Why it matters: Using the conversation context lets you turn earlier planning discussions into useful work-ready content without restating every detail. - Project-plan template components: Describe the main components included in the simplified project-plan template: project overview and objective, task list and timeline, milestones and dependencies, resource allocation, communication plan, and risk register.
Why it matters: Recognizing the template's sections helps you judge whether an AI-generated plan contains the practical information your project needs. - Template refinement and reuse: Explain that the generated template should be reviewed for missing or unsuitable sections, modified as needed, then saved as reusable actionable content for future projects before asking the chatbot to fill it with project details.
Why it matters: Reviewing and saving a good template turns a one-time chatbot response into a reliable starting point for future work. - Template export formats: State the available ways to retain chatbot-generated templates: copy and paste the content into a document, request a downloadable document when supported, or request markdown for documents and CSV for spreadsheets.
Why it matters: Knowing practical export options helps you keep and reuse AI-generated content even when download features are unavailable.
Key terms: template, project plan, project charter, meeting minutes, project closure report, Project Overview, Task List, Timeline, Milestones, Dependencies
5.1.4 Finalized output
This section shows how conversational project details can be used to have an AI fill a project-plan template, while recognizing when a more structured approach is needed.
- Conversational template completion: Describe how to provide a chatbot with project information in natural language and ask it to fill an existing project-plan template. Explain that the source information can be a write-up, bullet points, or spreadsheet rows, although this example supplies it conversationally.
Why it matters: You can turn project notes into a useful first-draft plan without manually completing every template field. - Migration plan derivation: Trace how the Oracle Database Migration description becomes a project plan with an objective, milestones, sequential dependencies, resource assignments, communications, and risks. State that the migration moves Oracle 19c to 23ai within a two-hour downtime window.
Why it matters: This example demonstrates how AI can transform scattered operational facts into standard project-management artifacts. - Draft refinement limits: Explain that AI can rapidly generate project-management suggestions, required documents, and a draft template output, but that the result still needs review and tweaks. Distinguish the convenience of natural-language prompting from its weakness when the plan requires many detailed, structured fields such as risk classifications and task dates.
Why it matters: Recognizing the boundary between a fast draft and a detailed plan helps you choose a prompt approach that produces usable results.
Key terms: template, natural language, chatbot, project plan, Oracle Database Migration, Oracle 19c, Oracle 23ai, milestones, dependencies, resource allocation
5.2 Problem formulation in action
This section demonstrates using top-down problem formulation to turn structured project data into a complete project-documentation template with a single precise AI prompt.
- Top-down problem formulation: Distinguish the top-down problem-formulation approach from the bottom-up approach used in section 5.1. Explain that problem formulation starts with a known desired outcome and planned steps, so an experienced practitioner guides the AI with a map rather than asking it to discover the path.
Why it matters: Starting from a defined outcome helps you get relevant results faster when you already understand the work. - Structured project template prompt: Describe how the example prompt creates project documentation from spreadsheet data. State that it assigns the AI an expert senior-project-manager role, specifies the required template sections, and instructs it to use dependency and milestone flags to populate their own sections rather than placing those flags in the task list.
Why it matters: Explicit output structure and transformation rules let AI convert operational data into usable documentation with fewer errors. - Expert-guided AI productivity: Explain why this is problem formulation rather than prompt engineering alone: the user tells the AI what to create instead of asking what to do or trying to learn from it. Describe how detailed structured input and preplanning reduce prompting effort, lower error risk, and make the AI a productivity tool whose output aligns with specific objectives.
Why it matters: Using your professional knowledge to define work precisely turns AI into a faster, more dependable assistant for workplace tasks.
Key terms: top-down approach, problem formulation, bottom-up approach, simplified template, Project Overview, Task List, Milestones, Dependencies, Resource Allocations, structured data inputs
5.3 Summary
This summary reinforces that context-rich, recursive prompts, sound problem formulation, and expert guidance produce more useful AI results.
Where this is used later
- Best practices for document handling with AI (chapter 6) · Recursive prompting in practice.
- Structuring and refining outputs (chapter 6) · Prompting iteratively in a practical workflow.
- Multiserver administration (chapter 9) · Iterating on prompts to refine an output.
- Format conversion (chapter 6) · Document format conversion builds on awareness that AI output can be requested in particular formats.
- Service desk therapy (chapter 8) · Producing finalized output from an iterative prompting workflow.
- Renaming files (chapter 9) · Producing a finalized, usable output from an AI prompting workflow
- Prompts used in this chapter (chapter 10) · the use of finalized output from an iterative prompting process
- Document summarization (chapter 6) · Document summarization requires specifying the desired result rather than asking the AI to determine the task.
Concepts explained in this chapter
- problem definitionProblem formulation is the process of defining and structuring a problem precisely so that it can be analyzed and solved effectively.
- project managementProject management is the practice of planning, coordinating, and controlling work to achieve defined objectives.
- prompt engineeringPrompt engineering is the deliberate design and refinement of instructions supplied to an AI system to improve the quality or relevance of its responses.
- conciseness criterionConciseness is the quality of communicating necessary information with minimal redundant or unnecessary wording.
- promptA prompt is an instruction or input supplied to an AI system to elicit a specified response or output.
- relevanceRelevance is the quality of directly relating information or requirements to the task or use case being addressed.
- zero-shot promptingZero-shot prompting is the technique of asking an AI model to perform a task without providing prior examples or task-specific context.
- project planA project plan is a document that specifies the tasks, responsibilities, schedule, milestones, dependencies, and resources for completing a project.
- recursive promptA recursive prompt is a prompting technique that incorporates an AI system's previous response into a subsequent prompt to develop a deeper or more nuanced interaction.
- ChatGPTChatGPT is OpenAI's conversational generative AI service for interacting with users through natural-language prompts and responses.
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.