14.2 How Project Managers Can Utilize GenAI

Integrating GenAI into project management offers powerful tools for automating tasks, enhancing decision-making, and generating project-related documents. GenAI is already transforming brainstorming sessions by allowing project managers to start conversations at a more advanced stage, saving time and boosting productivity^[1]. Beyond document generation and brainstorming, GenAI can support advanced use cases in project management. For instance, project managers can use GenAI for resource leveling, predictive analytics to forecast project delays or budget overruns, and automated reporting on project performance metrics. By tapping into these advanced capabilities, project managers can make more data-driven decisions and improve overall project efficiency.

Many project management software platforms have recently integrated AI features to enhance project planning, task management, and team collaboration. For instance, Atlassian Jira uses AI to streamline project tasks and provide context-aware suggestions to enable more efficient workflows. It also offers AI-driven analysis for issue tracking and decision-making through Atlassian Intelligence, enhancing task management with advanced automation features^[2][3]. Similarly, Microsoft Project has incorporated AI through its Copilot feature, which assists in project scheduling, resource allocation, and generating detailed status reports by summarizing data and suggesting actions. However, users should carefully review all AI-generated content for accuracy before relying on it^[4].

Project managers often utilize GenAI by creating prompts in LLMs (Large Language Models) such as ChatGPT, Microsoft Copilot, and Google Gemini. GenAI’s effectiveness largely depends on how well it is guided through prompt engineering. By providing well-structured and clear prompts, project managers can harness the full potential of these AI tools and ensure they receive useful and accurate outputs for their project needs.

In this section, we’ll explore key principles of prompt engineering and how to craft effective prompts to maximize GenAI’s capabilities in project management.

14.2.1 Prompt Engineering and Tips for Project Managers

Prompt engineering is creating and refining prompts to guide AI tools in producing useful responses. This skill is essential for project managers, as the quality of AI outputs is closely tied to how questions or instructions are framed. A well-structured prompt should be clear, specific, and contextual to the project. As AI continues to evolve, understanding the nuances of prompt engineering has become crucial for maximizing the effectiveness of tools like GPT^[5].

Key Tips for Effective Prompt Engineering:

1. Be Specific:
   • Instead of asking vague questions, you should structure the prompts with clear goals and context. For example, instead of asking, “Create a project plan,” you should try “Create a project plan for developing a mobile app, including task dependencies, deadlines, and resource allocations.”
   • Specificity helps the AI focus on the right elements. Providing precise details, such as expected outcomes or criteria, enables the model to generate more contextually relevant responses^[6]. Besides, being specific allows GenAI to jumpstart brainstorming, delivering a “solid strawman” of ideas that can be refined by the team, thereby saving time and enhancing the brainstorming process^[7].
2. Provide Context:
   • AI tools often perform better when given sufficient background information. If asking for risk analysis, you should include project-specific details such as scope, timeline, and key stakeholders. For example, “Analyze risks for a two-year e-commerce platform development project with a $5 million budget.” Contextual prompts provide more relevant and actionable insights. Adding context ensures that LLMs can leverage their knowledge to deliver more accurate and human-like outputs^[8].
3. Break Complex Tasks into Smaller Prompts:
   • If the task is complex, consider breaking it down into multiple steps. For instance, when creating a project charter, you could first ask for an outline and then prompt the AI to fill in sections like objectives, stakeholders, and deliverables. This approach improves the depth and clarity of the responses^[9].
4. Use Iterative Prompting:
   • Sometimes, the first output won’t be perfect, so you better iterate. You should refine the prompts based on the AI’s responses. You should adjust the prompt to request more detail or a different format if you receive an incomplete or vague answer. Iterative prompting allows the model to generate more precise results as it hones in on the desired outcome^[10].
5. Request Alternative Solutions:
   • You should ask the AI for multiple options or approaches to encourage creative problem-solving. For example, “Generate three alternative risk mitigation strategies for a delayed project launch.” This encourages the AI to think beyond the first answer and offer multiple perspectives. Providing examples or asking for variations in the solution allows the AI to better tailor the output to the user’s needs and deliver more flexible responses^[11].
6. Clarify Output Format:
   • You should specify how you want the output delivered—bullet points, a chart, or a detailed report. For example, “Provide a project schedule in Gantt chart format, listing task dependencies and milestones.” This improves the usability of the generated content and helps integrate it more effectively into project documentation.

By using these strategies, project managers can ensure that GenAI tools provide more accurate and useful results tailored to their project’s specific needs. Let’s explore how project managers can apply these principles in predictive project management and agile methodologies.

14.2.2 Predictive Approach

In predictive project management, project managers rely heavily on upfront planning, detailed documentation, and strict adherence to timelines and budgets^[12][13]. GenAI can assist in these areas by offering predictive insights, calculating key performance metrics and EVM (Earned Value Management) indicators, and generating detailed project charters, plans, and documents.

Below are examples of how project managers can leverage prompt engineering to utilize GenAI for predictive project management. The first example of “Creating a Project Charter” includes more detailed guidance on how project managers can interact with an LLM such as ChatGPT. Including iterative steps and instructions helps project managers refine and input the outputs.

14.2.2.1 Prompts for Creating a Project Charter

First, you should provide background information about the project. Attaching documents such as the needs assessment, business case, benefits management plan, SWOT analysis, agreements, enterprise environmental factors, and organizational process assets to the prompt will help the LLM generate a more accurate and contextually relevant project charter by allowing the AI to understand the project’s goals, constraints, risks, and stakeholders in greater detail. This ensures the AI considers all necessary factors while drafting, which can lead to a more comprehensive and tailored output.

Step 1: Initial Prompt and Output

You will receive a draft project charter after you provide the above prompt to the LLM. You should review the output for accuracy and completeness.

Step 2: Refining the Output

Now, it is time to assess the draft charter. You should identify areas needing more detail, such as specific deliverables, potential risks to the project, and key assumptions or constraints. For instance, while the initial project charter includes high-level objectives and key deliverables, it lacks clarity in areas such as the specific responsibilities of key stakeholders (e.g., project sponsor and project manager) and the potential risks and assumptions related to the project. You must recognize that including detailed responsibilities for stakeholders ensures accountability and clear communication during the project. Additionally, outlining risks such as hardware procurement delays or system downtime will help in proactive risk management, while clarifying assumptions (e.g., resource availability or vendor timelines) will ensure the project is grounded in realistic expectations.

Step 3: Improved Output

The LLM can now generate a more detailed project charter based on the refined prompt.

Step 4: Finalizing the Charter

Now, you can evaluate whether further iterations are necessary. You can refine it further by adding timeframes for the deliverables or specific performance metrics to track the project’s success.

14.2.2.2 Applying Prompt Engineering Tips to Project Planning and Forecasting

Prompt Example:
“Generate a detailed project plan for the construction of a new office building. The plan should include the following: key phases (e.g., design, foundation, construction, finishing), milestones for each phase, task dependencies (e.g., foundation must be completed before construction), and resource allocations (e.g., number of workers and materials required for each phase). The project should be completed within 18 months and stay within the $5 million budget.”

How This Prompt Relates to Key Tips for Effective Prompt Engineering:

1. Be Specific:
   This prompt avoids vague requests by specifying clear goals and context. Instead of simply asking for a general project plan, it provides exact requirements such as the project phases (design, foundation, construction), milestones, and resource allocations. Being specific ensures that the AI focuses on the right elements, enabling it to generate more contextually relevant responses. By including exact figures like the budget ($5 million) and the timeframe (18 months), the AI is guided to provide a more realistic and applicable project plan.
2. Provide Context:
   The prompt ensures that the AI is given sufficient context about the project’s scope, including constructing a building and constraints like the timeframe and budget. Providing background information (e.g., key project phases, budget, and dependencies) allows the AI to produce actionable insights more aligned with the actual project environment. Adding context ensures the AI delivers human-like outputs that are better aligned with the project’s needs.
3. Break Complex Tasks into Smaller Prompts:
   In this case, the prompt breaks down the complex task of creating a project plan by specifying multiple components, such as phases, milestones, dependencies, and resource allocations. Each part is a smaller, manageable task, improving the responses’ depth and clarity. Breaking down tasks helps the AI generate more accurate outputs for each component of the project plan.
4. Use Iterative Prompting:
   The initial output may not be perfect, so iterative prompting should be used. After receiving the first draft, you can refine the prompt by asking for more detail, such as specific resource breakdowns or contingencies for potential delays. This iterative process helps you guide the AI toward more precise and polished results over multiple iterations.
5. Request Alternative Solutions:
   You can further refine the project plan by asking the AI for alternative solutions if certain aspects of the project change. For example, you could ask, “Generate an alternative project plan if the budget is reduced by 20% and the timeline is extended by three months.” This encourages the AI to think creatively, offer multiple approaches, and help you see different strategies for managing project constraints. Requesting variations in the solution allows the AI to tailor its output to different scenarios.
6. Clarify Output Format:
   The prompt can specify the desired output format by requesting that the AI deliver the plan in a detailed list or Gantt chart format. For example, you could include an additional instruction: “Provide a project schedule in Gantt chart format, listing task dependencies and milestones for each phase.” This makes the AI-generated content more usable and ready for integration into the project documentation.

This enhanced prompt helps you understand how to apply effective prompt engineering techniques to create a well-structured project plan. It incorporates specificity, context, and iterative improvement while encouraging flexibility through alternative solutions and clarifying the output format. By following these guidelines, you can craft prompts that lead to more relevant, actionable AI-generated responses, ultimately helping you manage projects more effectively in your career.

14.2.2.3 Short Prompt Examples

Below are other examples of creating prompts in the predictive project management approach. It is important to remember that you should always provide the background information and relevant documents as input for the LLM to generate an effective response. These documents could include project charters, business cases, stakeholder lists, and risk assessments, which provide the necessary context for the LLM to create precise and actionable outputs. By supplying detailed and relevant information, you ensure the AI can generate results aligned with the project’s purpose and the triple constraint of scope, schedule, and cost. Additionally, refining and iterating the prompt based on initial outputs will help you improve the accuracy and relevance of the generated responses.

1. Forecasting Project Costs:
   • “Based on the following project details, estimate the total project cost, including labor, materials, and overheads. The project involves building a software application over 12 months with a team of 8 developers, 2 designers, and 1 project manager. Assume average salaries are $100,000 for developers, $85,000 for designers, and $120,000 for the project manager. Include additional costs for software licenses ($10,000) and cloud infrastructure ($20,000).”
   • This prompt details labor costs, resource counts, and overheads, ensuring the AI generates a cost estimate grounded in realistic assumptions. Providing specific salaries and additional costs helps the AI give a more precise budget forecast.
2. Risk Identification and Mitigation:
   • “Create a risk management plan for launching a new product in an emerging market. Consider potential risks, including market volatility, regulatory changes, supply chain disruptions, and local competition. Suggest a mitigation strategy for each risk, including short-term and long-term actions.”
   • This prompt provides the AI with a detailed project context (a product launch in an emerging market) and specific risks to focus on, ensuring a comprehensive risk management plan. It asks for both short-term and long-term mitigation strategies, helping you see how risk management should evolve.
3. Calculate Earned Value Management (EVM) Metrics:
   • “Based on the project’s budget and percentage of completion, calculate the Earned Value (EV), Planned Value (PV), Cost Performance Index (CPI), and Schedule Performance Index (SPI). The project budget is $500,000 and is currently 50% complete, with actual costs of $300,000. Analyze these metrics to assess whether the project is on track.”
   • This prompt specifies key project metrics (EV, PV, CPI, SPI) and provides the project’s financial details, ensuring the AI can generate accurate EVM calculations. By including actual and planned costs, the prompt helps the AI precisely analyze whether the project is on budget and schedule.
4. Create a Gantt Chart with Dependencies:
   • “Generate a Gantt chart for the development of a new software application. The chart should include tasks such as initial planning, software design, coding, testing, and deployment. Show dependencies between tasks (e.g., coding must be completed before testing), and include a timeline with estimated start and end dates for each task.”
   • This prompt clearly defines the tasks and dependencies involved in the software development project. By specifying key phases and dependencies, the AI can generate a Gantt chart that visually organizes the timeline, making it easier for project managers to track progress and ensure tasks are completed in the correct order.
5. Perform Critical Path Analysis:
   • “Identify the critical path for a construction project involving building a new office complex. The project consists of tasks such as obtaining permits, laying the foundation, constructing the frame, and interior finishing. Provide the durations for each task and indicate which tasks are on the critical path.”
   • This prompt specifies the tasks and requires the AI to calculate task durations and dependencies. By asking for the critical path, the prompt ensures the AI identifies the longest sequence of tasks with zero slack that must be completed on time to avoid project delays, providing key insights for project scheduling.
6. Estimate Resource Allocation:
   • “Estimate the resource allocation for a 12-month software development project, including developers, testers, and designers. The project requires 5 developers, 3 testers, and 2 designers. Allocate resources based on a $1 million budget and ensure the project stays within these limits while meeting deadlines.”
   • This prompt details resource requirements and budget constraints, ensuring the AI generates a precise resource allocation plan. By specifying roles and budget, the prompt allows the AI to deliver an allocation plan that optimizes resources and ensures the project remains on track financially.
7. Monitor Project Performance:
   • “Generate a performance report for the first quarter of a two-year project to build a new distribution center. The project has a budget of $20 million and is currently 25% complete, with actual costs of $6 million. Provide insights into cost performance, schedule performance, and any risks or delays that might impact the project timeline.”
   • This prompt specifies key performance metrics such as cost and schedule, ensuring the AI can generate a report evaluating the project’s current status. The prompt helps the AI provide accurate insights into the project’s health by including actual costs and completion rates.
8. Predict Delays and Offer Contingency Plans:
   • “Predict potential delays for a large-scale construction project to build a new office park. The project faces supply chain issues, labor shortages, and regulatory approval delays. For each predicted delay, provide a contingency plan (e.g., alternative suppliers, subcontractors, or fast-tracking approval processes).”
   • This prompt outlines specific issues that could cause delays, allowing the AI to predict potential disruptions to the project. By requesting contingency plans, the prompt ensures that AI generates proactive solutions to mitigate these delays and keep the project on track.

14.2.3 Adaptive Approach

In agile project management, flexibility and continuous adaptation are key, as project managers prioritize responding to change over following a rigid plan. Agile methodologies focus on iterative development, collaboration, and delivering incremental value throughout the project lifecycle^[14][15]. In this approach, pivoting based on feedback and evolving requirements is crucial to success. GenAI can significantly support agile projects by helping create user stories, estimate story points, generate agile-specific artifacts such as burndown charts, and provide insights during retrospectives. By leveraging GenAI’s capacity for real-time data analysis and dynamic output generation, project managers can make informed decisions, adapt quickly, and streamline processes for better collaboration and efficiency.

Below are examples of how project managers can use prompt engineering to utilize GenAI effectively in agile project management, along with iterative instructions to guide interaction with AI models like ChatGPT.

14.2.3.1 Applying Prompt Engineering Tips to Scrum Project Story Estimation

Prompt Example:
“Estimate the story points for the following user stories in a Scrum project to develop an e-commerce website. The user stories include ‘As a customer, I want to be able to filter products by category,’ ‘As an admin, I want to manage product inventory from the dashboard,’ and ‘As a customer, I want to receive an email confirmation after placing an order.’ Consider complexity, effort, and uncertainty in your estimations, and provide a brief rationale for each story point assignment. Deliver the output as a table with user story descriptions and corresponding story points.”

How This Prompt Follows the Key Tips for Effective Prompt Engineering:

1. Be Specific:
   The prompt includes specific user stories, clarifying what the AI should focus on for estimation. Details such as “filter products by category” and “manage product inventory” guide the AI in estimating more accurately.
2. Provide Context:
   The prompt mentions that the estimations are for an e-commerce website within a Scrum project, providing context on the environment and type of project and helping the AI tailor its responses appropriately.
3. Break Complex Tasks into Smaller Prompts:
   The prompt focuses on estimating story points for specific user stories rather than tackling all aspects of the project at once. This ensures the task is manageable, allowing for more focused, accurate responses.
4. Use Iterative Prompting:
   If the initial output is too vague or incomplete, you could refine the prompt by asking the AI to consider specific criteria (e.g., effort required for each task or potential blockers), thus improving the accuracy of the estimations.
5. Request Alternative Solutions:
   You could iterate further by asking for alternative estimations based on different team sizes or sprint lengths, encouraging AI to provide multiple perspectives on effectively completing the user stories.
6. Clarify Output Format:
   The prompt specifies that the output should be delivered in a table format with user story descriptions and story points, ensuring the AI generates a clear, organized response that is easy to use in the Scrum planning process.

14.2.3.2 Further Prompts for the Scrum Project Story Estimation

First, it’s important to provide relevant background information about the project. Attaching documents such as user story lists, project backlog, sprint planning documents, and velocity reports will help the LLM generate more accurate estimates. These documents provide context about the team’s capacity, the complexity of tasks, and the project’s goals, which allows the LLM to offer more contextually relevant estimates. The AI can then estimate the effort required for each user story, consider dependencies, and account for the team’s velocity when assigning story points.

Step 1: Initial Prompt and Output

After providing the above prompt to the LLM, you will receive an initial set of story point estimates. Review the output for accuracy, consistency with previous sprints, and whether the rationale provided makes sense in the context of the project’s complexity and team capacity.

Step 2: Refining the Output

Now, it is time to assess the initial estimates. You should identify areas that might need further refinement, such as complexity assumptions, dependencies on other stories, or constraints in team velocity. For instance, while the initial estimates offer a basic understanding of complexity, they might lack the necessary detail regarding uncertainties, technical challenges, or team capacity. Clarifying dependencies and additional complexities can help refine these estimates.

Refining the Prompt:

“Refine the estimates for the user stories by considering potential technical challenges, such as integrating with third-party services (e.g., email API for confirmation emails) and ensuring the filter function interacts with the entire product catalog. Also, adjust for any dependencies on other user stories in the backlog. Provide updated estimates with a brief explanation of how these factors impact the story points.”

Step 3: Improved Output

The LLM can now generate refined estimates based on the additional context about dependencies, technical challenges, and team capacity.

Step 4: Finalizing the Estimations

Now, you can evaluate whether further iterations are necessary. You could refine the estimates further by incorporating team velocity or estimating based on additional factors like unforeseen risks (e.g., bugs in the API integration). For example, you can ask for alternative solutions based on different sprint lengths or team sizes.

14.2.3.3 Short Prompt Examples

1. Create User Stories for a Mobile App Development Project:
   • “Generate five user stories for a Scrum project to develop a mobile app for an online bookstore. Include user roles and specific functionality, such as browsing books, making a purchase, and writing reviews.”
   • This prompt helps project managers quickly generate clear and actionable user stories, ensuring the team understands the key functionalities from the user’s perspective. It provides a structured format to guide development.
2. Prioritize Tasks for a Sprint:
   • “Prioritize the following tasks for the first sprint of a project to develop a mobile app for online shopping. Tasks include setting up the user authentication system, designing the homepage, integrating the payment gateway, and creating the product catalog. Prioritize based on business value, technical dependencies, and team capacity.”
   • This prompt helps agile teams prioritize tasks based on key factors like business value, dependencies, and capacity, ensuring the most critical work is completed first. It supports effective sprint planning by guiding teams in organizing their work to deliver value early.
3. Generate a Sprint Backlog:
   • “Generate a sprint backlog for the first sprint in a project to develop an online learning platform. Include the top priority user stories and break them down into tasks, such as designing the course interface, setting up user accounts, and creating a payment system.”
   • This prompt helps teams plan and prioritize their work for the sprint, breaking down user stories into smaller, actionable tasks. It aids in sprint planning and task assignment.
4. Create a Retrospective Discussion Guide:
   • “Create a discussion guide for the sprint retrospective of a project that faced challenges in completing tasks on time. Include prompts to encourage team members to share their concerns, reflect on what went well, and propose solutions for improving communication and collaboration in the next sprint.”
   • This prompt focuses on power skills, such as leadership and communication, by encouraging a culture of continuous improvement. It guides the team in reflecting on successes and challenges while fostering open communication and problem-solving in agile environments.
5. Create a Burndown Chart:
   • “Generate a burndown chart for a Scrum team’s current sprint, with 10 user stories and a sprint length of two weeks. Show how much work remains and track progress over time.”
   • This prompt allows the team to visualize their progress throughout the sprint, providing real-time insights into whether they are on track to complete the work. It supports ongoing sprint monitoring and helps teams adjust priorities if needed.