Prompt Engineering for PMs: AI Guide

Let me share a personal experience with you. Just six months ago, I took three hours to write a project status report, while now, with the help of ChatGPT, I do it in 12 minutes. And no, this is no exaggeration. The only difference is that I learned to 'talk' with AI.

Some of you might be thinking that Artificial Intelligence is replacing project managers. On the contrary, it is enhancing our work. The differentiator is that we need to know the correct questions to ask. This is the role of prompt engineering. If you are doing PMP certification training, acquiring AI skills at the same time will give you a significant edge over your peers.

Prompt engineering is a skill, and that skill is selecting the right wording for the AI to give you the best results. Think of it as learning a new dialect with extreme levels of specificity.

Take the difference between "help me with my project plan" and "please develop a work breakdown structure for a 6-month duration software development project with 8 members focusing on mobile app deployment". The second prompt is contextual and sets constraints. As a result, it offers a suitable answer.

Most project managers I know interact with AI tools in a relaxed manner. They ask simple questions and expect to obtain reasonable responses. That method ignores the real potential that these tools provide. Professional prompt engineering changes AI from being a sophisticated search engine to becoming a true business ally.

After some trial and error with AI tools, I know what actually works for project management:

ChatGPT is my first choice for completing everyday tasks. Meeting agendas, status reports, and risk assessments from ChatGPT work great. I rely on Claude when I need to analyze thick project briefs or when I need a project document to be deconstructed. I rely on Gemini for tasks that require up-to-the-minute data.

When you add AI capabilities to traditional PM skills, the benefits you get from project management increase significantly. You are not choosing between human expertise and artificial intelligence. You are combining both.

Elements of a Good Prompt

A prompt is comprised of four main components: the role, the task, contextual information, and the expected output. I will illustrate these components with the following example.

Weak prompt: "Provide a timeline for the project."

Strong prompt: "As an expert project manager, please outline a timeline for the project, focused on the customer portal, that spans 12 weeks. The working team consists of 2 developers, 1 designer, and 1 QA tester. Please include milestones, dependencies, and structure the plan as a Gantt chart."

Observe the differences. The strong prompt explains the role of the person (expert PM), what he must achieve (timeline), gives background (composition of the team and how long they will work), and tells how he must present the result (Gantt chart).

Chain-of-Thought Prompting

Complicated choices require reasoning in a step-by-step manner. I do this when I need to select a certain project selection method.

Instead of asking, "Should we go for Project A or Project B?" I ask, "Please compare Project A and Project B in terms of 1. the resources required, 2. the potential for ROI, 3. alignment to strategy, 4. risks involved, 5. feasibility in terms of timelines, and then make a decision, explaining your reasoning for each one."

This way, I make sure the AI is considering each of the elements one by one. This increases the quality of the result significantly.

Planning and Documentation

I keep a library of reusable prompts. Here's one I use weekly:

"Draft an agenda for a project kickoff meeting for a [project type] with these stakeholders: [list]. Allocate time for each topic and specify discussion and decision-making outcomes. Provide the information in a table with columns for Time, Topic, Owner, and Expected Outcome."

This template helps me save 45 minutes each week.

Risk Management

Everyone working in project management knows that there are typical project risks associated with certain types of projects. The challenge lies in adapting these for your projects. Risk analysis prompts can be constructed in the following way.

"Analyze this project scope: [paste scope]. List 10 possible risks and assign each of the following categories: technical, resource, schedule, and stakeholder. For each risk, assign a response of high, medium, or low, an impact assessment, and describe two risk mitigation strategies."

This output is a baseline, and I supplement it with feedback from the project team.

Stakeholder Communication

Spending time on making the same message fit different customers is time-consuming, and AI helps overcome this. I often use:

"Rewrite this project update for three audiences: 1) Executive leadership (focus on business impact and ROI), 2) Technical team (emphasize implementation details), 3) End users (highlight benefits and timeline). Each version must be 200 words or fewer."

This way, in a matter of minutes,s I create three different versions for three different audiences.

Optimizing Context Windows

Each AI tool has its own token limits. For example, ChatGPT has a limit of about 8,000 words worth of context. During larger project management assignments, I divide info into manageable chunks.

Instead of putting an entire project charter, I can structure my prompts for the AI. I provide the background first, followed by the objectives, then the scope, and finally finish with the constraints. Keeping prompts in this sequence helps the AI maintain focus, while I abide by the limitations.

Automating Workflows Using Prompt Chaining

Chained prompts can be useful for completing tasks that contain multiple steps. For example, here is my workflow for resource allocation:

Prompt 1: "List all tasks in this project management plan with estimated hours."

Prompt 2 (using the output from the first prompt): "Assign these tasks to team members based on the following skill profiles: [paste profiles]."

Prompt 3: "Describe any scheduling constraints and provide possible solutions."

Each of these prompts builds on the output from the previous one. As a result, the output is all-encompassing and takes into account multiple variables that I might overlook.

To evaluate the success of AI integration, examine the following metrics.

Time savings: I noted a 35% decrease in the time I spend on documentation due to the constant use of AI prompts.

Quality improvements: Initial drafts are of much higher quality, leading to fewer revision requests on deliverables.

Team satisfaction: Administrative tasks are completed much quicker, which my team appreciates, leaving them more time for higher-level strategic work.

Many professionals underscore the importance of artificial intelligence integration within digital tools when enhancing skills through PMP certification training, as it often proves to be a significant factor in a competitive job market.

Dependence on AI answers: Always fact-check AI information, especially if important decisions depend on it. Sometimes, systems create fictitious versions of facts. I learned, for example, when a system created a fictitious project management methodology that sounded credible but didn't exist. It invented a project management methodology that sounded credible but didn't exist.

Artificial intelligence systems are designed to help users with their tasks, but you often need to adjust a system in order to get the most optimal results.

Integrate project management AI systems by using individual prompts in the categories of project plan, project execution, project monitoring, project closing, and stakeholder management.

Set a goal of creating three templates this week that will help automate recurring tasks. Write them down. Test them. Improve them. Within your team, share the templates that work best. After a month, you will have a collection of templates that will give you a few hours of free time each week.

Managing a project's key performance indicators shows you the extent to which AI integration improves your performance metrics, and by extension, the time you will save.

The evolution of technology, specifically Artificial Intelligence, is happening at record-shattering speeds. With all the advancements happening in the further development of Project Management tools, Human Project Managers must retain the ability to apply critical thinking. This is the hallmark of superior project management.

The best way to start is to integrate tools that help you complete monotonous tasks. Spend time experimenting, and create a detailed prompt. Refine and adjust that prompt until you are satisfied with the results. Rinse and repeat. This is how you will manage your tasks and help you improve your projects to levels you never thought were possible.