30 AI Prompts for Project Managers
Copy-ready prompts that assume you will attach the requirements, the notes or the project data — built around PM frameworks (WBS, RAID, RACI, pre-mortem) so planning and reporting take minutes.
Each prompt gives the AI a delivery-manager persona, references what you will attach, and applies a project framework with a defined deliverable. These prompts are tool-agnostic and optimised to work across all major AI assistants — Claude, Microsoft Copilot and ChatGPT (and Google Gemini). Replace the [bracketed placeholders] with your own details, and always review AI output before you act on it.
Planning and scoping
Work Breakdown Structure Builder
Scoping a new project before estimating or staffing it
project goal description
- Break the goal into phases with milestones
- List work packages under each milestone, one deliverable per package
- Keep packages mutually exclusive and collectively exhaustive
- Flag any package with unclear scope before estimating
Work breakdown structure with milestones.
See example AI output
Phase 1: Discovery (Milestone: Requirements Signed Off) - WP1.1 Stakeholder interviews leads to Deliverable: interview notes - WP1.2 Requirements doc leads to Deliverable: signed-off BRD Phase 2: Build (Milestone: MVP Complete) - WP2.1 Core module dev leads to Deliverable: working MVP - WP2.2 QA cycle leads to Deliverable: test report [Flag] WP2.1 scope for the reporting module is unclear - confirm with sponsor before estimating.
Project Charter Draft
Getting a new project formally approved and scoped
project name
- State the objective and success criteria
- Define what's explicitly in and out of scope
- List key stakeholders, assumptions and constraints
- Keep it to one page with unambiguous scope boundaries
One-page project charter.
See example AI output
Project: Warehouse Inventory App Objective: cut stock-count errors by 50% within Q4. In scope: barcode scanning, live stock dashboard. Out of scope: supplier ordering module. Stakeholders: Ops Director (sponsor), Warehouse Leads, IT. Success criteria: error rate under 2%, live by 30 Nov. Assumptions: handheld scanners already owned. Constraints: RM80k budget, 10-week timeline.
Prioritised Backlog From Requirements
Turning raw requirements into a sprint-ready backlog
requirements list
- Convert each requirement into a user story
- Write testable, independent acceptance criteria per item
- Prioritise by value versus effort
Prioritised backlog of user stories with acceptance criteria.
See example AI output
1. As a shopper, I want saved payment methods so checkout is faster. AC: card saved after first purchase; removable in settings. Value: High, Effort: Low. 2. As a shopper, I want order tracking so I know delivery status. AC: status updates within 1hr of carrier scan. Value: High, Effort: Medium. 3. As an admin, I want a returns dashboard so refunds are faster. Value: Medium, Effort: Medium.
Task Estimate And Critical Path
Sequencing tasks and spotting the critical path before committing a schedule
task list
- Estimate relative effort per task, stating the basis of estimate
- Sequence tasks and map dependencies
- Flag the critical path and mark low-confidence items
Sequenced task list with dependencies and critical path.
See example AI output
1. Design API (3d, basis: similar past sprint) leads to blocks Task 3 2. Set up test env (2d) leads to parallel 3. Build endpoints (5d, low confidence - scope unclear) leads to depends on Task 1 4. Integration testing (3d) leads to depends on Task 3 Critical path: 1 leads to 3 leads to 4 (11 days). Task 2 has slack.
Project Timeline Outline
Laying out a phased schedule with protected milestone dates
project name and duration
- Lay out phases across the duration with milestone target dates
- Note dependencies between phases
- Add buffer around critical milestones, not the whole schedule
Phased timeline with milestone dates and buffer.
See example AI output
Project: CRM Migration (12 weeks) Phase 1 Discovery (Wk1-2) - Milestone: Data map signed off (Wk2) Phase 2 Migration build (Wk3-7) - Milestone: Test migration complete (Wk7), 3-day buffer Phase 3 UAT (Wk8-10) - Milestone: Sign-off (Wk10) Phase 4 Go-live (Wk11-12) - Milestone: Cutover (Wk12), depends on UAT sign-off.
Status and reporting
Weekly Status Report Builder
Sending the regular weekly status update to sponsors
this week's project updates
- Summarise progress against plan
- State an honest RAG status with the reason, no green-washing
- List risks, issues and decisions needed
- Outline next steps, exceptions only
Weekly status report.
See example AI output
Project: Mobile App Revamp - Week 14 Status: AMBER (reason: design sign-off 4 days late) Progress: 60% of sprint 6 complete, on track for feature freeze. Risks: app store review may slip launch by 1 week. Decisions needed: approve revised launch date of 15 Aug. Next steps: finalise design by Fri, start regression testing Mon.
RAG Status One-Liner
Giving a fast, honest health check when someone asks how it's going
project notes + project name
- Assign a defensible RAG colour
- Justify it in one line
- State the single most important fact for the reader
One-line RAG status with justification.
See example AI output
Project: Payroll System Upgrade - AMBER Justification: two of five test cycles failed due to tax-table errors. Most important: go-live on 1 Sept is at risk unless fixes are verified by 20 Aug.
Executive One-Liner And Summary
Briefing an executive who only has 30 seconds
no file - name the project
- Lead with a one-line status headline
- State current position and the key risk
- State the specific ask
One-liner plus a three-sentence summary.
See example AI output
Headline: on track, but vendor delay threatens the Sept launch. Summary: development is 85% complete and testing starts next week. The hardware vendor has slipped delivery by 10 days, which could push go-live past the board deadline. We need approval to expedite-ship at an extra RM12k to protect the date.
Audience-First Stakeholder Update
Tailoring the same project update to a specific stakeholder group
project data
- Lead with what matters most to this specific audience
- Summarise progress and flag risks
- State the decision needed
Stakeholder update, audience-first.
See example AI output
For the Finance Committee - Budget System Rollout What matters to you: spend is currently 4% under budget. Progress: 3 of 4 modules live, final module in UAT. Risks: currency-conversion bug found, fix in progress. Decision needed: approve 1-week extension to finish regression testing before sign-off.
Delay Explanation Message
Telling stakeholders a milestone has slipped, with a credible recovery plan
project name + delay details
- State what slipped and why, candidly and without blame
- Explain the downstream impact
- Present a credible recovery plan with a revised date
Delay-explanation message.
See example AI output
Subject: update on Website Relaunch timeline The content migration ran two weeks behind schedule due to a larger-than-expected volume of legacy pages needing rework. This pushes the relaunch from 1 Aug to 15 Aug. Recovery plan: adding one contractor and running migration and QA in parallel from next week. Revised go-live: 15 August. We will confirm progress at Friday's checkpoint.
Risk and issues
RAID Log Builder
Standing up a full risk, assumption, issue and dependency log for a project
project context description
- List risks with likelihood, impact and mitigation, owned
- List assumptions and dependencies
- List issues with owners, prioritised
RAID log covering risks, assumptions, issues and dependencies.
See example AI output
Risks: vendor API instability (Likelihood: Med, Impact: High, Mitigation: build fallback queue) - Owner: Tech Lead Assumptions: client will provide test data by Wk3 Issues: staging environment down since Mon - Owner: DevOps, due Wed Dependencies: legal review must complete before the vendor contract is signed.
Single Risk-Register Entry
Logging one specific risk with full detail for governance review
name of the specific risk
- Describe the risk and its trigger
- Rate likelihood and impact, quantified where possible
- List current controls
- State response strategy and owner
Single risk-register entry.
See example AI output
Risk: key developer resignation mid-sprint Trigger: notice period started Likelihood: Low (15%) | Impact: High (2-week slip) Current controls: pair programming on critical modules Response: mitigate - cross-train a second engineer within 2 weeks Owner: Engineering Manager
Issue Triage Action Log
Working through a pile of open issues and unblocking the critical path first
list of open issues
- Prioritise issues by impact and urgency
- Unblock the critical path first
- Assign owners and due dates to each
Prioritised issue action log.
See example AI output
1. Payment gateway timeout (blocks launch) - Owner: Backend Lead, Due: Thu 2. Broken analytics tracking - Owner: Data team, Due: Fri 3. Typo in confirmation email - Owner: Content, Due: next sprint Critical path unblocked once Item 1 is resolved.
Project Pre-Mortem
Stress-testing a project plan before it kicks off
no file - name the project
- Assume the project has already failed
- List failure reasons ranked by probability, brutally honest
- Give a preventive action for each
Pre-mortem findings with preventive actions.
See example AI output
Project: New Branch Opening (assume failed) 1. (40%) Staff not hired in time leads to Prevention: start recruitment now, not at Wk8 2. (30%) Landlord fit-out delays leads to Prevention: lock a penalty clause into the lease 3. (20%) Local permits rejected leads to Prevention: submit the permit application this week, not next month
Issue Resolution Options
Deciding how to fix a live issue with clear trade-offs
issue details
- Summarise the issue
- Propose three resolution options
- Quantify cost, time and risk for each
- Give a clear recommendation
Resolution options with a recommendation.
See example AI output
Issue: database performance degraded 40% after last release. Option A - Roll back: cost RM0, time 1hr, risk: loses new features. Option B - Add indexing: cost RM5k, time 2 days, risk: low, does not fully fix. Option C - Re-architect query layer: cost RM40k, time 3 weeks, risk: high but permanent fix. Recommendation: Option A now to stabilise, then Option C planned for next quarter.
Stakeholder communication
Stakeholder Map And Comms Plan
Setting up who to talk to, about what, and how often on a new project
no file - name the project
- List stakeholder groups and rate interest and influence for each
- Note what each group cares about
- Set communication approach and cadence, prioritising high-influence high-interest
Stakeholder map with a communications plan.
See example AI output
Project: ERP Rollout Finance Director - Influence: High, Interest: High. Cares about cost control. Comms: weekly 1:1. End users (Ops team) - Influence: Low, Interest: High. Cares about ease of use. Comms: bi-weekly newsletter. IT Security - Influence: High, Interest: Med. Cares about compliance. Comms: monthly steering update.
Project Kickoff Email
Launching a new project team with clear roles and first actions
no file - name the project
- State the objective and scope
- List roles and the timeline
- Set ways of working
- Give the first action so people can move today
Kickoff email.
See example AI output
Subject: kicking off Project Nova - let's build! Objective: launch the loyalty app by 1 Oct. Scope: iOS/Android app plus points engine; excludes in-store kiosks. Roles: Aisyah (PM), Ravi (Tech Lead), Mei (Design). Timeline: 8 weeks, sprint reviews every Friday. Ways of working: async updates in Slack, stand-up 9:15am daily. First action: everyone reviews the requirements doc by Wed.
Expectation-Reset Message
Proactively flagging a scope or timeline change before stakeholders notice
scope or timeline change description
- State the change and the reason honestly
- Describe the impact
- Present options with a recommendation, solution-oriented not just bad news
Expectation-reset message.
See example AI output
Subject: change to the Analytics Dashboard scope The change: real-time data feed is being descoped from Phase 1. Reason: source system API is not ready until Q1 next year. Impact: dashboard will show data with a 24-hour lag at launch. Options: (1) launch with daily refresh now, add real-time later, or (2) delay launch 3 months. Recommendation: Option 1 - ship value now, upgrade later.
Technical-To-Business Translator
Turning an engineering update into something a business sponsor can act on
technical update
- Read the technical update
- Extract the business impact and the decision needed
- Rewrite in plain language, no jargon
Business-friendly update.
See example AI output
Technical note: API rate-limiting caused 502 errors during peak load, mitigated via a caching layer. Business version: some customers saw the app fail to load during our busiest hour last week. We fixed it by storing frequently used data closer to users. No customer data was affected, and we recommend no further action at this time.
Steering Committee Talking Points
Prepping for a steering committee where decisions must get made
no file - name the project
- Summarise current status
- List decisions needed with a recommendation for each
- List risks to escalate, decision-focused not status-focused
Steering-committee talking points.
See example AI output
Project: Data Centre Migration Status: 70% complete, on budget. Decision needed: approve additional RM50k for backup hardware - Recommend: approve, risk of data loss otherwise. Risk to escalate: sole networking vendor has a 3-week lead time - Recommend: approve dual-sourcing now. Next milestone: cutover rehearsal, 2 Aug.
Meetings and coordination
Meeting Agenda Designer
Designing a focused, time-boxed agenda before a project meeting
meeting type + project name
- State the meeting objective and required pre-read
- Time-box each agenda item realistically
- Define the outcome to leave with
Time-boxed meeting agenda.
See example AI output
Sprint Planning - Project Falcon (60 min) Objective: commit to Sprint 9 scope. Pre-read: backlog groomed items, capacity sheet. 0-10min: review last sprint's velocity 10-40min: walk backlog, size top items 40-55min: confirm sprint goal and commitments 55-60min: risks and blockers Outcome: agreed sprint backlog with owners assigned.
Meeting Notes To Actions
Converting raw meeting notes into tracked decisions and actions
raw meeting notes
- Extract decisions made
- Extract actions with owners and due dates, unambiguous ownership
- Extract open questions, keep nothing lost
Three sections: decisions, actions, open questions.
See example AI output
Decisions: proceed with Vendor B for hosting. Actions: - Send contract to Legal - Owner: Priya, Due: Wed - Update budget sheet - Owner: Farhan, Due: Fri Open questions: does Vendor B support our SSO requirement? Need an answer before the contract is signed.
Stand-up Summary Builder
Rolling up individual daily updates into a team summary with blockers surfaced
team's individual updates
- Surface blockers first
- Summarise progress across the team
- Identify who needs help today
Stand-up summary.
See example AI output
Blockers: Ravi blocked on staging access (needs IT ticket approved). Progress: login flow done and merged; checkout UI 80% complete; QA started regression on payments. Needs help: Mei needs a design review slot today to unblock checkout.
Acceptance Criteria Writer
Writing testable acceptance criteria before a feature goes into a sprint
feature or task description
- Write acceptance criteria for the feature
- Use given-when-then format where it fits
- Ensure every criterion is unambiguous and verifiable
Set of acceptance criteria.
See example AI output
Feature: password reset via email Given a registered user requests reset, When they submit a valid email, Then a reset link is sent within 60 seconds. Given the link is older than 30 minutes, When clicked, Then the user sees 'link expired' and can request a new one. Given an invalid email is submitted, When the form is submitted, Then a generic message is shown with no account enumeration.
Overdue Action Chase Messages
Chasing multiple people for overdue actions without sounding heavy-handed
list of people + overdue actions
- Draft one message per person
- Reference the specific overdue action by name
- Set a firm but friendly deadline, not heavy-handed
One chase message per person.
See example AI output
To Aisyah: Hi Aisyah, checking in on the vendor comparison sheet due last Friday - could you share it by end of day tomorrow so we can finalise the choice at Thursday's review? To Daniel: Hi Daniel, the test cases for the payments module were due Monday. Can we get them by Wed morning so QA is not blocked?
Retrospectives and improvement
Retrospective Session Designer
Planning a sprint or project retro that produces real commitments
no file - name the project/sprint
- Choose a retro format
- Write questions covering what went well, what didn't, what to change
- Design how findings become owned commitments, psychologically safe
Retrospective plan with questions.
See example AI output
Format: Start-Stop-Continue, 45 min, anonymous input first. Questions: what should we start doing? What should we stop? What should we continue? Safety note: no names attached to raised issues, facilitator reads items aloud. Commitments: top 3 voted items become actions with an owner and a check-in date at the next retro.
Retro Feedback Theme Synthesiser
Turning raw retro feedback into themes and concrete actions
raw retro feedback
- Cluster raw feedback into themes
- Quantify how often each theme appeared, theme over anecdote
- Produce three concrete improvement actions with owners
Themes with counts, plus three improvement actions.
See example AI output
Theme 1: unclear requirements (mentioned 6x) Theme 2: too many meetings (mentioned 4x) Theme 3: slow code review turnaround (mentioned 5x) Actions: 1. Add a requirements checklist before sprint start - Owner: PM 2. Cut recurring meetings from 5 to 3 per week - Owner: Team Lead 3. Set a 24-hour SLA for code reviews - Owner: Eng Lead
Lessons-Learned Documentation
Capturing project learnings for the next team to reuse
no file - name the project
- Document what worked, concretely
- Document what to avoid
- Give specific, reusable recommendations for future teams
Lessons-learned documentation.
See example AI output
Project: Office Relocation What worked: early vendor walk-throughs caught 3 issues before signing the lease. What to avoid: do not finalise furniture orders before IT confirms the desk-power layout - caused rework. Recommendations: build a 2-week buffer before any move date; assign one single point of contact for building management liaison from day one.
Delivery Slip Root-Cause Analysis
Digging into why a milestone slipped, beyond the surface symptom
slipped milestone description + context
- Analyse why the milestone slipped
- Identify systemic causes, not just symptoms, no blame
- Recommend process changes to the system
Root-cause analysis with recommended process changes.
See example AI output
Milestone: UAT sign-off slipped 3 weeks. Symptom: testers found requirements ambiguous. Systemic cause: no formal requirements review gate before build started. Another cause: test environment provisioned only after build finished, compressing test time. Process changes: add a requirements review gate before Sprint 1; provision test environments in parallel with build, not after.
Improvement Ideas Action Plan
Prioritising a backlog of process-improvement ideas into a sequenced plan
list of improvement ideas
- Assess effort and expected benefit per idea
- Assign owners to each
- Sequence quick wins first, realistic pacing
Prioritised improvement action plan.
See example AI output
1. Standardise sprint templates (Effort: Low, Benefit: Med) - Owner: PM - Do first 2. Automate status report generation (Effort: Med, Benefit: High) - Owner: PMO - Next 3. Restructure the team into two pods (Effort: High, Benefit: High) - Owner: Head of Delivery - Plan for next quarter
More prompt packs by function
View the full Prompt LibraryFrequently Asked Questions
Each applies a project framework — WBS for scoping, RAID for risk, pre-mortem for foresight, RACI for accountability — and assumes you will attach the requirements, notes or project data. That turns hours of documentation and reporting into minutes, so you focus on the coordination and judgement that keep projects on track.
Yes — the prompts work from your attached requirements, status notes and project data. Use approved enterprise AI for anything confidential, and anonymise client or commercial details where you can.
Copilot is convenient inside Microsoft Project, Teams and Planner; Claude and ChatGPT are strong for planning, summarising and drafting. Many PMs also connect AI to their PM tools. These prompts work across all of them.
It drafts a solid first-pass plan, work breakdown and risk log from your inputs, which you refine with real-world judgement. Treat it as an accelerator, not autopilot.
Yes. AITraining2U's courses are HRD Corp SBL-KHAS claimable for eligible Malaysian employers.
Yes. Give the AI your task or work-package list plus the team roles, and ask it to assign Responsible, Accountable, Consulted and Informed for each item, then flag any task with no owner or more than one Accountable. This surfaces accountability gaps fast, a common cause of stalled decisions. It works well combined with a stakeholder map: use the RACI for day-to-day task ownership and the stakeholder map for communication cadence. Review the AI's first draft with your team, since only they know the true reporting lines and informal influence a matrix alone will not capture.
AI will not replace the conversation, but it removes the admin around it. Before planning, ask it to size and sequence a groomed backlog from your notes. During the sprint, feed it daily updates and it drafts the stand-up summary with blockers surfaced first. After the sprint, paste retro feedback and it clusters it into themes with suggested actions in seconds instead of the usual sticky-note exercise. The team still discusses and decides; AI just handles the synthesis and formatting so ceremonies spend more time on judgement and less on paperwork.
Increasingly, yes. Copilot is built into Microsoft Project, Planner and Teams. Claude and ChatGPT can connect to Jira, Asana, Trello and similar tools through their connector or plugin ecosystems, letting the AI read tickets and draft updates without you exporting data manually. If your organisation has not enabled a connector yet, the fallback is simple: export or copy the relevant board, backlog or issue list and paste it into the prompt. Either way, always check your company's data-handling policy before connecting AI to a tool that holds client or commercial information.
Treat AI the same way you would treat an external contractor with no NDA by default. Use your organisation's approved enterprise AI tool where chats are not used for training and data stays within agreed boundaries, rather than free consumer accounts. Strip or replace client names, contract values and personal data with placeholders before pasting anything in, and only attach the minimum data needed to get a useful answer. For genuinely sensitive material, such as unsigned contracts, personal data or security details, keep it out of any AI tool entirely, enterprise or not, unless your legal and IT teams have specifically cleared that use case.
Yes. Give it your team roster, their current allocations and the upcoming task list, and ask it to flag over-allocated people, identify who has spare capacity, and suggest a rebalanced plan. It is also useful for scenario-testing: ask what happens to the timeline if one person is pulled onto another project for two weeks. The AI is only as good as the data you feed it, so keep allocation numbers current, and always sanity-check its suggestions against real skills and availability, since a spreadsheet does not know who is on leave or who is the only person who understands the legacy system.
Go beyond prompts — train your team
Prompts are the start. AITraining2U runs hands-on, HRD Corp SBL-KHAS claimable AI training for Malaysian teams — from everyday AI productivity to building agents that run project management workflows end-to-end.