This chapter moves from theory to practice, showing how AI chatbots can help IT professionals manage projects by crafting better prompts. It demonstrates a two-part approach: starting with basic prompt engineering to generate quick, work-ready outputs, then advancing to problem formulation to direct the model more precisely. The overarching goal is to use AI not just as a teacher for learning best practices, but as a tool for creating tangible project artifacts efficiently.
In the first part, the chapter illustrates how a vague, zero-shot prompt yields generic advice, while a slightly more contextual prompt improves relevance and actionability. It introduces the principles of clarity, conciseness, relevance, and specificity, then applies recursive prompting to build on chat context without repeating details. This leads to identifying necessary documents, generating a simplified project template, and populating it with an Oracle database migration example—producing a project plan, milestones, dependencies, a communication plan, and a risk register in minutes. The lesson: natural language prompts can rapidly deliver usable content, but become cumbersome when detailed data (like dates and risk ratings) are required.
The second part shifts to a top-down problem formulation approach: the prompter tells the AI exactly what to produce and supplies structured inputs (for example, spreadsheet data). By directing the model to assemble specified sections and infer milestones and dependencies from flags, the output becomes more complete and precise, with fewer iterations and less room for error. The chapter concludes that combining domain expertise with strategic prompt crafting and structured inputs transforms AI into a powerful productivity tool—yielding faster, more relevant project management deliverables tailored to real-world IT needs.
Summary
Effective prompt engineering leads to better AI interactions in project management.
Detailed context in prompts yields more actionable AI responses.
Recursive prompts utilize previous outputs for coherent, advanced queries.
Problem formulation enables quicker, more relevant results from AI models.
Expert knowledge in project management enhances AI productivity and output quality.
FAQ
What’s the difference between prompt engineering and problem formulation in this chapter?Prompt engineering is the bottom-up approach: you start with broad or basic prompts and iteratively refine them to get useful output. Problem formulation is the top-down approach: you decide the exact outcome and structure you want, then direct the AI to produce it, often with structured inputs (like a spreadsheet). The chapter shows both methods, with problem formulation reducing back-and-forth by telling the AI exactly what to do.Why did the first “How can I be a good project manager?” prompt underperform?It was a zero-shot prompt with minimal context, so the AI produced generic, wide-ranging advice. Without clarity on scope, role, and desired deliverables, the model “rambling” is likely. Adding concrete context and constraints focuses the answer on actionable outputs.How can I quickly improve a basic project management prompt?State your role, scope, and constraints, and ask for the specific type of help you need. For example: “I’m leading a small IT project with no budget tracking. Which tasks should I complete?” This boosts clarity, conciseness, relevance, and specificity—the four prompt principles used in the chapter.What are recursive prompts and when should I use them?Recursive prompts build on the existing chat context, letting you ask follow-ups without restating details. Use them to sequence work, like first asking for tasks, then for documents, then for templates based on prior answers. Note that starting a new chat clears context.Which documents should I create for a small IT project?Core documents highlighted include: Project Charter, Project Plan, Communication Plan, Risk Register, Task Tracker, Meeting Minutes, and a Project Closure Report. You can tailor the set to your needs—exclude what’s not needed and deepen the ones you’ll use.How do I get the AI to generate a practical, work-ready template?Ask for a simplified template with the sections you need (for example: Project Overview, Task List, Milestones, Dependencies, Resource Allocation, Communication Plan, Risk Register) and specify exclusions. You can also request a specific output format (Markdown for documents, CSV for tables) for easy reuse.What’s the best way to provide data so the AI fills the template accurately?Use structured inputs when possible (for example, paste rows from a spreadsheet). Direct the model explicitly—tell it to include the provided data in the template and to use dependency/milestone flags to build those sections without duplicating the flags in the task list. This reduces errors and increases precision.How do I tailor AI output to my project constraints (like no budget tracking)?State constraints up front and instruct the model to exclude irrelevant sections. For instance: “I’m not tracking budget; exclude charter, meeting minutes, and closure report—use the other suggestions.” The chapter shows the model honoring such limits and returning only what’s useful.How do I apply the four prompt principles effectively?Clarity: say exactly what you’re doing (small IT project, your role). Conciseness: include only necessary details. Relevance: align the ask to your scenario (not generic PM advice). Specificity: state constraints, desired sections, and format so the output is immediately usable.Any tips for exporting and reusing AI-generated content?Ask for Markdown to copy/paste documents without a paid plan, and CSV for spreadsheets like task lists or risk registers. Save useful templates for future projects, and keep them accessible to your team. Some chat tools can also generate downloadable files, though that may be a premium feature.
pro $24.99 per month
access to all Manning books, MEAPs, liveVideos, liveProjects, and audiobooks!
AI for Everyday IT ebook for free
