Free resource · Word Kit

AI Kit for Project Managers: 24 use cases and their prompts, in three levels

From meeting minutes to portfolio co-pilot: ready-to-use prompts to help your project teams and PMO develop their skills on their own, step by step.

📄 Word, 44 pages🧭 24 cards, 3 levels🗓️ 6-week course🆓 Free

Download the kitSee the 24 use cases

Most project managers in large organizations have a generative AI tool at their fingertips: Copilot in Teams, a ChatGPT or Claude enterprise account, or Gemini in Workspace. What's lacking isn't the tool itself, but rather a process of continuous improvement. Everyone tries it on their own, gets a rough report or a generic schedule, and concludes that "it doesn't work for project management." Meanwhile, a minority integrates AI into their routines and saves several hours a week.

This page explains what recent surveys show about the use of AI by project teams, why a tiered progression is better than a list of prompts, what the 24 use cases in the kit contain, and gives you the complete kit, in Word format, to get started today on your real projects.

1. AI in project teams: what the numbers show

The PMI has been measuring the gap since 2023. That year, 21% of surveyed project professionals were using AI in their work, 49% reported little or no experience with it, and 91% believed it would transform the profession. In March 2025, the study "The Project Professional's GenAI Journey" distinguished two profiles: "Explorers," who had experimented a little, and "Trailblazers," who had integrated AI into their practices. 48% of the latter saw significant productivity gains, compared to 3% of the former; 58% of Trailblazers regularly experimented with new uses, compared to 12% of Explorers. The difference lies not in talent or the tool itself, but in consistency and methodology.

Team expectations remain consistent across surveys. Capterra, in July 2025, surveyed 2,545 professionals in eleven countries, including France: automating administrative tasks (48%), anticipating problems (37%), and better managing risks (28%). When an AI function is available in the project tool, 79% of teams use it daily. Wellingtone prioritizes automated progress reports, suggestions for next steps, and early warning signals. Gartner predicted as early as 2019 that 80% of current project management tasks would be handled by AI by 2030.

The real issue is the level of support. In Wellingtone's survey of PMOs in July 2026, 86% of organizations were using Microsoft 365 and 73% provided access to Copilot; only 14% had implemented a structured skills development program. The tool is there, but the progress isn't. That's precisely what this kit aims to address.

2. Why a progression by level, and not just a list of more prompts?

There's no shortage of prompt libraries for project managers: Asana, Smartsheet, Google Workspace, Glean, and a dozen others publish them. They share three common flaws. They're simplistic: forty prompts presented in no particular order, leaving beginners unsure where to start. They stop at the AI ​​output: nothing on what to prepare beforehand, nothing on what to check afterward. And they ignore the context of a large organization: confidential data, AI policy, meeting transcripts, governance bodies, and the PMO.

The kit takes a different approach to the problem. It follows the three ways of working with AI described in the AI ​​Fluency framework published by Anthropic in 2025: automation, where the AI ​​performs a specific task on command; augmentation, where humans and AI build together in multiple rounds; and agency, where humans configure the AI ​​to then perform tasks on their behalf. These three modes yield three natural levels, and the four skills of the same framework (delegate, describe, discern, and be diligent) provide the self-assessment grid. Each level covers the eight project management scenarios, from scoping to portfolio management, so the kit can also be used on demand: you find your scenario, then the level where you feel most comfortable.

Each fact sheet finally explains what the lists don't: when to use the case, what to prepare, what to expect, the three points to check before sharing, the variations depending on the tool, and the next fact sheet. It's this last link that makes the progression concrete: each fact sheet clearly indicates where it leads.

3. The three levels and the six-week course

Level	Working method	What you know how to do at the end	Indicative duration
Level 1 · Start	The AI ​​performs one task at a time, based on what you give it.	Get in minutes a report, an agenda, a task list, an initial risk list, a flash report or an email that you can distribute after reviewing.	2 weeks, 20 to 30 minutes per day
Level 2 · Structuring	Co-constructing AI: you provide your documents, you engage in dialogue over several rounds, you make it work in stages.	Produce complete project deliverables that can be presented to the committee: scoping note, reverse planning and RACI, listed risk register, variance analysis, audience reporting, closing balance sheet, portfolio review.	2 weeks, 30 to 45 minutes per day
Level 3 · Systematize	You configure the AI ​​to work for you and the team: standing instructions, projects and agents, prompt chains.	Having an assistant who knows your project, a weekly reporting chain to action plan, a dashboard, a pre-mortem, a knowledge base and a portfolio co-pilot.	2 weeks, 45 minutes a day, then a few minutes a week

The proposed learning path takes six weeks: four worksheets per week, presented in the order of the situations, and a validation challenge at the end of each level, based on a real-world project. At level 1, participants produce and distribute the minutes of their next meeting. At level 2, they present a deliverable created with AI to a committee. At level 3, they activate their project assistant and have a colleague use it. The twelve-question self-assessment in the user manual indicates the entry level; the progress log and self-assessment grid in the appendices measure the progress made.

4. The 24 use cases, from scoping to portfolio

Eight situations, three levels, one sheet per box. The situations are based on those we work on in the AI training for project managers and PMOs, supplemented by the closing and the portfolio.

Situation	Level 1 · Start	Level 2 · Structuring	Level 3 · Systematize
Feasibility study	Rephrase a vague objective into a measurable one and list the questions to ask the sponsor	Draft the scoping document after a mock interview with the sponsor, and map the stakeholders	Create your own permanent project assistant (ChatGPT project or Claude, Copilot agent, Gem Gemini)
Planning	Break down a deliverable into tasks (first WBS)	Build the reverse timeline with dependencies and milestones, then the RACI matrix	Reschedule by scenario and export to Planner, Project or Excel
Risks and resources	Establish an initial list of risks	To categorize the risk register, build mitigation plans, and analyze the workload	Pre-mortem and "red team": having your plan attacked by AI
Meetings	Transforming raw notes into minutes, preparing an agenda	Preparing a steering committee: support, expected decisions, difficult questions	The chain "from the report to the updated action plan", and the committee simulation
Monitoring and reporting	Write the weekly news flash report using a few bullet points.	Analyze the discrepancies between the schedule and budget and propose a corrective action plan.	Dashboard: define the indicators, generate the model, automate the update
Communication	Write a progress or follow-up email in three tones	Adapt the same report for each audience and write the executive summary.	Produce the project's communication kit: messages for each population group, FAQs, calendar
Closure and capitalization	Prepare questions for a retrospective or review	Prepare the closing report and lessons learned summary based on the meeting minutes.	Create a project knowledge base that can be reused by the team
Portfolio and PMO	Consolidate several progress points into a portfolio summary	Preparing a portfolio review: criteria, scenarios, trade-offs	The portfolio co-pilot: weak signals, alerts, team prompt library

The details of each case follow: the situation, what you get, and the point to check before using the result. The complete prompt is in the kit.

Level 1 · Start: AI executes

1.1 Reformulating a vague objective into a measurable objective

Framing. Time saved: 30 minutes per objective. At launch, when the sponsor has expressed an intention ("modernize the request management tool") without success criteria, and before the scoping meeting where you will have to clarify it.

What you get: one objective in one sentence, with an indicator and a target indicated [to be confirmed] if it is missing; two alternatives; eight ordered questions that structure your next exchange with the sponsor. To check : the proposed target values ​​are well reported [to be confirmed] and not invented.

1.2 Decompose a deliverable into tasks (first WBS)

Planning. Time saved: 1 hour per work batch. When you need to build the schedule and the task list doesn't yet exist; to arrive at the planning workshop with the team equipped with a basis to correct rather than a blank page.

What you get: a table of 15 to 40 lines, with a verifiable result per task, and a list of frequent omissions (validations, recipe, communication, training, data recovery, production deployment). To check : Each task has a verifiable result: "organizing meetings" is not a task.

1.3 Establish an initial list of risks

Risks and resources. Time saved: 45 minutes per workshop. Before the risk workshop, to provide it with a basis to challenge; at each phase change; after an event that changes the context (new supplier, reorganization).

What you get: fifteen risks formulated as "if... then...", classified into five families, with an early warning sign for each, and three underestimated risks with supporting arguments. To check : The risks are specific to your project: reformulate those that remain generic (“lack of resources”).

1.4 Transforming raw notes into minutes, preparing an agenda

Meetings. Time saved: 30 to 45 minutes per meeting. After every meeting of which you have taken notes, even telegraphic ones; before a meeting of which you know the objective but not yet the agenda.

What you get: a one-page report, ready to be distributed after review, and a timed agenda where each item has a nature and a person responsible. To check : Each action was actually said in the meeting: AI readily fills in the gaps.

1.5 Write the weekly flash report using a few bullet points

Monitoring and reporting. Time saved: 30 minutes per week. Every week, when you have the facts in mind but no time to format them, and the report ends up being sent late or too long.

What you get: a 150-word flash report, readable in one minute, the last part of which clearly states what the reader must decide and by when. To check : The status and its reason are yours: AI should not soften an orange or a red. Its prompt is reproduced below.

1.6 Write a progress or follow-up email in three tones

Communication. Time saved: 15 minutes per email. Following up with a late contributor, informing the sponsor of a new fact, responding to a service provider, requesting arbitration: every time the tone counts as much as the content.

What you get: three emails ready to send, with subject line, where you choose the tone based on the relationship and history. To check : The commitments mentioned exist (dates, reports): check before sending the firm version.

1.7 Prepare questions for a retrospective or review

Closing and summarizing. Time saved: 30 minutes per retrospective. End of phase, end of sprint, end of project, or after an incident, when you need to lead a reflection session and you don't have a framework.

What you get: a timed schedule, twelve open questions, a format for commitments, and three facilitation pitfalls to avoid. To check : The questions are open-ended: no "is it that".

1.8 Consolidate several progress points into a portfolio summary

Portfolio and PMO. Time saved: 1 hour per consolidation. If you are a PMO or program manager, you receive five to twenty heterogeneous reports and you need to produce an overview for management.

What you get: a summary table, three points of attention with supporting arguments, a list of factual inconsistencies and one question per project manager. To check : The statuses are those declared by the project managers, not those that the AI ​​would have "corrected".

Level 2 · Structuring: Co-constructed AI

2.1 Draft the scoping document after a mock interview with the sponsor, and map the stakeholders

Framing. Time saved: 3 hours per framing note. The first formal deliverable of the project, when the brief received is thin and you want to arrive in front of the sponsor with a document already structured and the right questions.

What you get: a three-page scoping note in which each hypothesis is marked, and a stakeholder map with an engagement action for each actor. To check : Each [hypothesis] is confirmed or removed before publication.

2.2 Construct the reverse schedule with dependencies and milestones, then the RACI matrix

Planning. Time saved: 2 hours per plan. When the end date is imposed and you have to show if it is feasible, then distribute responsibilities without any grey areas.

What you get: a reverse schedule with critical path and margins, two reasoned options if the date is untenable, a RACI matrix and its list of anomalies. To check : The dates are not binding until those doing the work have validated them.

2.3 Assess the risk register, develop mitigation plans, and analyze the workload

Risks and resources. Time saved: 2 hours per risk review. After the risk workshop, to move from a list to a rated and monitored register; at each monthly review; when an overload is looming.

What you get: a rated and classified register, five complete mitigation plans, a load analysis with critical people and three arbitrations. To check : The quotations are a collective decision: those of the AI ​​are only a starting proposal.

2.4 Preparing a steering committee: support, expected decisions, difficult questions

Meetings. Time saved: 2 hours per committee. One week before each committee meeting, when you need to get decisions and not just inform.

What you get: a support plan in messages, a sheet per decision with the cost of not deciding, ten difficult questions prepared. To check : The decision options are real and encrypted by you, not by AI.

2.5 Analyze the discrepancies between the schedule and budget and propose a corrective action plan

Monitoring and reporting. Time saved: 1 hour 30 minutes per analysis. As soon as the actual results deviate from the planned results, before the committee asks you to; at each milestone.

What you get: the discrepancies and a projection whose calculation is visible, a punctual / structural distinction, five classified actions and a ready alert message. To check : Redo the key calculations: the models make mistakes in arithmetic, hence the verifiable table.

2.6 Develop the same report for each audience and write the executive summary

Communication. Time saved: 1 hour per report. When the same progress report has to be read by the Executive Committee, the business units, the IT department, the service providers and the team, and everyone does not have the same time or the same rights.

What you get: four calibrated versions and a five-sentence executive summary, with the same numbers throughout. To check : The service provider version contains nothing internal: budget, tensions, other suppliers.

2.7 Prepare the closing report and lessons learned from the meeting minutes

Closing and capitalization. Time saved: 3 hours per balance sheet. At the end, when the documents exist but no one has the time to reread them to learn from them.

What you get: a four-page summary in which every claim is sourced, and ten actionable lessons. To check : Each statement refers to an actual document: if in doubt, ask for the citation.

2.8 Prepare a portfolio review: criteria, scenarios, trade-offs

Portfolio and PMO. Time saved: 3 hours per review. Before each portfolio review, when new demands arrive and capacity cannot keep up.

What you get: a marked and justified grid, three contrasting scenarios, five closed arbitration questions. To check : The notes are proposals to be reviewed by a committee, not a ranking.

Level 3 · Systematize: AI works for you

3.1 Create your permanent project assistant

Framing (cross-functional). Time saved: 2 hours per week thereafter. As soon as you find yourself repeating the same context at every prompt. The project assistant is the foundation for the other seven level 3 worksheets.

What you get: an assistant that responds with your context, formats and rules, in which the level 1 and 2 worksheets launch without further explanation. To check : The instructions and attached files do not contain any data that your charter prohibits from being stored in the tool.

3.2 Reschedule by scenarios and export to Planner, Project or Excel

Planning. Time saved: 2 hours per rescheduling. When an event disrupts the schedule: supplier delay, absence of a key resource, change of scope, and the sponsor wants options, not a statement of facts.

What you get: three alternative schedules that can be re-imported, with their cost, their risks and the question to ask the sponsor; a reasoned recommendation. To check : Re-import each scenario into your tool and check the dependencies: AI sometimes breaks one.

3.3 Pre-mortem and "red team": having your plan attacked by the AI

Risks and resources. Time saved: 1 hour 30 minutes, and surprises avoided. Before finalizing a plan or passing a commitment milestone; when everyone agrees, which is precisely the moment when blind spots are no longer visible.

What you get: ten dated causes of failure with their early warning signs, fifteen reasoned objections and the response or modification to be made to each. To check : Report the new causes in the risk register (sheet 2.3) with an owner.

3.4 The "from report to updated action plan" chain, and the committee simulation

Meetings. Time saved: 1 hour per week. Each week, after the progress meeting: the report, the action plan, and the follow-ups are done in three messages, always the same. Before each committee meeting: the rehearsal.

What you get: an up-to-date action plan where nothing has been lost, a dissemination message, five follow-ups ready; a committee rehearsal that reveals your weaknesses before the real session. To check : The reference action plan remains yours: version it, and the AI ​​will send back a copy.

3.5 Dashboard: define the indicators, generate the model, automate the update

Monitoring and reporting. Time saved: 2 hours per month, and reliable indicators. When monthly reporting relies on a file that only you know how to update, or when current indicators do not trigger any decisions.

What you get: one sheet per indicator with thresholds and limits, a three-sheet Excel template, a five-line monthly comment. To check : Check the formulas manually over a full month before trusting the model.

3.6 Produce the project communication kit: messages for each population group, FAQs, calendar

Communication. Time saved: one day per kit. When the project changes the work of entire populations and communication must be consistent over six months, carried by managers who have not followed the project.

What you get: a position line, a message per population in two versions, a FAQ of fifteen questions, a calendar aligned with milestones, three indicators. To check : Nothing is promised that the project cannot succeed: validation by the sponsor.

3.7 Establish a project knowledge base that can be reused by the team

Closure and capitalization. Time saved: hours on each new project. When the team or PMO has closed several projects and the lessons learned lie dormant in reports that no one reviews.

What you get: a deduplicated and sourced lesson database, and an assistant who provides it on demand, at the start of each project. To check : The documents are cleaned of all personal data before ingestion.

3.8 The portfolio co-pilot: weak signals, alerts, team prompt library

Portfolio and PMO. Time saved: 3 hours per week for a PMO. When the PMO consolidates weekly reports in a stable format and wants to see deviations before they are declared; when the team has accumulated prompts to harmonize.

What you get: Each week: a consolidation, sourced weak signals, one question per signal, three meeting topics. Once: a consistent prompt library for the team. To check : Weak signals are hypotheses: discuss them with the project manager before reporting them.

5. What a record contains, with an example

Each sheet fits on one page and follows the same format: a heading (situation, level, compatible tools, approximate time saved), "When to use it," "What to prepare," the prompt to copy with the sections to replace highlighted, "What you should achieve," "Check before using," "Variations depending on the tool" (Copilot in Teams, Outlook, Word, or Excel; ChatGPT, Claude, Gemini, Vibe), and "Next step." Here, to give you an idea, is the prompt for sheet 1.5, the weekly flash report.

And the three checkpoints that accompany it: the status and its reason are yours; the AI ​​must not soften an orange or a red; the expected decisions name the decision-maker and the date; nothing has been added to your bullet points. It is this discipline, repeated across 24 cards, that transforms an attempt into competence.

6. The five safety rules that accompany each prompt

A project manager handles sensitive data: budgets, supplier names, meeting transcripts, workload estimates per person. The toolkit's user guide sets out five rules, to be applied in addition to your organization's AI charter, which should never replace it.

• The authorized tool, for the data it authorizes. Internal data goes into a company tool; confidential, personal or classified data never goes into a free, consumer-grade tool.
• Anonymize before pasting. Names of people, clients and suppliers, sensitive amounts: labels are all that's needed for the prompt to work.
• AI proposes, the project manager decides. No date, no budget, no commitment produced by AI is released without human validation; arbitration, sensitive announcements and the evaluation of people cannot be delegated.
• Check figures, names, dates and references. The models invent with confidence; the checklist in Appendix C takes two minutes.
• Be transparent and keep track. Mention the use of AI when the charter requires it, and retain prompts and exits that were used in a decision. Since February 2, 2025, Article 4 of the European regulation on AI requires employers to take measures to develop the AI ​​skills of those who use it; Omnibus Regulation (EU) 2026/1744 has made this an obligation of means ("promote the development of AI skills"): the toolkit contributes to this.

7. What the kit contains

The kit is a 44-page Word document, with the sections to be replaced highlighted in yellow in each prompt. It opens with instructions: why this kit, the three levels and the six-week learning path, the twelve-question self-assessment, the prompt template with four ingredients (objective, context, sources, expectations) and the verification reflex, the five safety rules, and the key for the prompt sheets. Next come the three levels, eight sheets each, each level beginning with what you will be able to do and ending with a validation challenge and its passing criteria. The fourth part is addressed to the project team and the PMO: the "Monday prompt" ritual, the shared prompt library, the eight team usage rules, the five indicators for measuring gains, and the role of the PMO. Six annexes complete the kit: the use case matrix and the "which tool for what" table, the self-assessment grid by level on the four skills, the checklist for verifying an AI output, the progress log, the glossary and the sources.

8. Frequently Asked Questions

Who is this kit for?

For project managers, project directors, and PMOs in large organizations, regardless of their starting level: the twelve-question self-assessment indicates where to begin. Transformation and training departments can distribute it as is, as a self-learning course.

Is training required before using the kit?

No. The kit is designed for self-sufficiency: each sheet explains what to prepare, provides the prompt to copy, and lists the points to check. Training on your own files accelerates skills development and addresses the specific needs of your organization, but it is not a prerequisite.

What tools does it use?

With Microsoft 365 Copilot, ChatGPT, Claude, Gemini and Vibe from Mistral (formerly Le Chat). The prompts are neutral; each card indicates useful alternatives when the tool changes the situation, for example Copilot in Teams for a transcribed meeting, Copilot in Excel for a dashboard, or a Claude Project for a large file.

How to use it in a team or at the PMO?

The fourth part of the kit offers five lightweight tools: the fifteen-minute "Monday prompt" ritual, a shared library of prompts, eight team usage rules, five indicators to measure gains, and the role of the PMO as the tool carrier, in connection with ambassador networks and managers.

Are the prompts safe for confidential project data?

The kit establishes five security rules: the tool must be authorized for the data it authorizes, anonymization must be performed before pasting, human validation of all dates, budgets, and commitments, systematic verification of figures, names, and references, and transparency and traceability. Each sheet also indicates the sensitive data specific to its use case.

How long does the journey take?

Six weeks, 20 to 45 minutes per day, two weeks per level, with a validation challenge at the end of each level. The kit can also be used on demand: the matrix in Appendix A allows you to directly find the sheet that corresponds to your current situation.

9. Going further

The kit allows you to progress independently; it does not replace a workshop on your own projects. AI training for project managers and PMOs NextStart.AI's 1.5-hour or half-day training program gets your teams working on their scoping notes, committees, and real-world reports, with 70% hands-on practice, and they leave with the prompts and agents built during the session. The level 3 worksheets also link to the resources of the library app.nextstart.ai, including the AI ​​co-pilot of project portfolios and the preparation of a portfolio review.

Read also : Manager's toolkit for dealing with AI: getting your team on board, AI Ambassador Network: The 2026 toolkit for building your collective, AI Charter in Business: The 2026 Model (Downloadable), AI Adoption Roadmap in 12 Months: The 2026 Toolkit et Measuring the uses, value and ROI of AI: the 2026 toolkit.

The entire series of free kits. To drive adoption: AI charter template, AI training plan template, AI Ambassadors Kit, Manager's toolkit for dealing with AI, Adoption Roadmap Kit, AI maturity grid, AI Measurement and ROI Kit, AI Governance Kit, Gift Sets Cowork, Copilot Studio kit, Claude Skills kit et Vibe coding kitFor the AI ​​project sponsor: AI Act kit, GDPR kit, Security and data kit, Shadow AI kit, AI and Social Dialogue Kit, frugal AI kit, AI Sovereignty Kit et responsible AI kitBy profession: AI HR kit, AI marketing kit, AI finance kit, AI kit for lawyers, AI kit for salespeople, AI innovator kit et communicator's AI kit.

10. Sources

The figures and references cited on this page and in the kit are based on the following sources, consulted in September 2026.

• Project Management Institute, The Project Professional's GenAI JourneyMarch 2025.
• Project Management Institute, Shaping the Future of Project Management With AI, October 2023.
• Capterra, AI for Project Managers Is Reshaping PM Software Decisions, survey from July 2025 of 2,545 professionals in 11 countries.
• Wellington, AI Is Arriving in Project Management, But Are PMOs Ready?July 2026, and The State of Project Management, annual report.
• Gartner, AI will assume 80% of project management tasks by 2030press release dated March 20, 2019.
• Rick Dakan, Joseph Feller and Anthropic, AI Fluency Framework, 2025 (CC BY-NC-SA 4.0 license).
• Microsoft, Get started writing prompts in Microsoft 365 Copilot, and Copilot Scenario Library: Execute a project review.
• Asana, Prompt AI project management: the complete and practical guide Smartsheet, AI Prompts for Project Managers Google Workspace AI Prompts for Project Managers.
• Regulation (EU) 2024/1689 of 13 June 2024 laying down harmonised rules on artificial intelligence, Article 4, Mastering AIapplicable from February 2, 2025.