AI Prompts for Project Managers: The Free Prompt Library
Why a prompt library matters
The single biggest predictor of whether a PM gets consistent value from AI is whether they use a prompt library — a versioned collection of proven prompts they reuse every week — or whether they improvise.
PMs who improvise get wildly inconsistent output. PMs who lock their prompts get a status report in 15 minutes every Friday, week after week, without variance.
This article shares the core prompts senior PMs use in real enterprise programs. Copy, adapt, and add to your library.
The seven core prompts every PM should have
1. Weekly status report
```
You are a senior program manager writing a weekly status report for
an executive steering committee.
RULES:
1. Use ONLY the source material provided below. Do NOT invent facts.
2. If a data point is missing, insert "PM to confirm" — never guess.
3. Match the attached template EXACTLY, including headings and order.
4. Keep the Executive Summary to 4 sentences maximum. Plain business
English, no marketing language.
5. RAG rules: Green = on track; Amber = slippage < 10% or risk needing
mitigation; Red = slippage >= 10% or unresolved blocker.
TEMPLATE:
[paste your standard weekly status template]
SOURCE MATERIAL:
[paste Jira exports, workstream lead updates, meeting transcripts,
current RAID log, and last week's report]
```
2. Meeting summary + action items
```
Read the meeting transcript below. Produce:
(a) A one-paragraph narrative summary (5 sentences max).
(b) A structured list of decisions with owners named.
(c) A structured list of action items with owners, due dates (if
stated), and status (Open).
(d) A follow-up email draft (3 short paragraphs) to send to attendees.
Do NOT include filler discussion. Focus on decisions and actions.
TRANSCRIPT:
[paste transcript]
```
3. Risk / RAID log review
```
You are a senior risk manager reviewing the following artifacts:
- Project charter
- Latest weekly status report
- Current RAID log
Identify:
1. Risks that appear stale (>60 days old, unchanged) and should be
re-baselined or closed.
2. Risks that appear to have been closed prematurely.
3. New risks implied by the status report but NOT in the RAID log.
Provide each finding with: (a) which item, (b) why it's stale/
premature/missing, and (c) recommended next step.
ARTIFACTS:
[paste all three]
```
4. Stakeholder-tailored communications
```
Below is a single source of truth (this week's project status).
Produce THREE versions of this update:
(a) CEO version: 2 sentences maximum. Strategic tone.
(b) Steering committee version: full RAG format, ~200 words.
(c) Team version: plain English, ~150 words, explaining the "why"
behind key decisions.
Do not repeat information verbatim across versions — adapt tone and
detail to each audience.
SOURCE:
[paste weekly status]
```
5. Executive narrative for board/steering
```
Write a two-page steering committee narrative using the source
material below.
Structure:
1. Strategic frame (2 sentences: why this program matters right now)
2. Portfolio RAG summary (table format)
3. The three things steering needs to know this cycle
4. Specific asks: decisions requested from steering
Tone: confident, tight, no marketing language, no jargon.
Length: 400–500 words max.
SOURCE:
[paste portfolio status report, key project narratives, current asks]
```
6. Requirements review / gap analysis
```
Read the following requirements document. As a senior PM, identify:
1. Contradictions between requirements (list each with the two
conflicting requirements verbatim).
2. Requirements that lack acceptance criteria.
3. Requirements that appear to imply significant scope not called out
elsewhere.
4. Missing requirements: capabilities the target system is likely to
need but which are not in the document.
For each finding, quote the relevant text and explain your reasoning.
REQUIREMENTS:
[paste requirements doc]
```
7. Change request impact assessment
```
Below is a change request. Produce a first-draft impact assessment
covering:
1. Scope impact (what changes)
2. Schedule impact (best/expected/worst case)
3. Cost impact (budget line items affected)
4. Risk impact (new risks or changes to existing risks)
5. Recommendation: approve, reject, or approve-with-conditions
Rules: Be specific. If you don't have enough information to estimate
something, say so — do not guess.
CHANGE REQUEST:
[paste CR]
BASELINE PLAN:
[paste relevant sections of baseline]
```
Six operating principles for your prompt library
1. Version it. Keep the library in a shared document (Notion, Confluence, OneNote) with a "last updated" date on each prompt.
2. Test-drive new prompts before making them canonical. Run a candidate prompt for 2 weeks before adding it to the library.
3. Never share the raw source material in prompts to consumer AI tools. Use your organization's enterprise AI tenant only.
4. Share the library across your team. PMOs where every PM uses the same status report prompt produce dramatically more consistent output.
5. Iterate quarterly. Models change. Your organization matures. Review the library every 90 days.
6. Always leave a human review step. The AI produces the draft. The PM signs the output.
Common mistakes with prompts
- Vague prompts. "Write a status report" gets a generic result. Specific, rule-bound prompts (like the ones above) get consistent output.
- Forgetting to include the template. AI cannot match your PMO's format if you don't paste it in.
- Skipping the "do not invent" rule. Without an explicit constraint, models will fill in plausible-sounding but incorrect facts.
- Using consumer AI accounts for confidential data. Don't. See our full breakdown: How to Use AI for Project Status Reports.
The expanded library
The seven prompts above are the foundation. A full enterprise-ready prompt library includes 30+ prompts across:
- Kickoff artifacts (charter, WBS, communication plan)
- Vendor RFP scoring
- Cross-project dependency analysis
- Executive one-pagers
- PMO governance documents
If you want the full curated library — with real enterprise case studies and the governance framework — Claude PM Pro is the 12-module program built for senior PMs leading enterprise delivery.
Or start with the free 60-minute masterclass to see the framework in action.
Frequently Asked Questions
What is a project management AI prompt library?
A prompt library is a versioned collection of proven AI prompts a PM reuses week to week for repetitive workflows: status reports, meeting summaries, risk analysis, stakeholder communications, and executive narratives. Locking your prompts is the #1 factor that separates PMs who get consistent AI results from those who don't.
Why do I need a prompt library instead of improvising?
Two reasons: (1) output quality is dramatically more consistent when you use the same prompt every time, so you can trust the format; and (2) time savings compound — a good prompt shaves 30–60 minutes off a task, and you get that saving every single time you reuse it. PMs who improvise their prompts each week get inconsistent output and lose most of the ROI.
Should I share my prompt library with my team?
Yes. A shared prompt library is one of the highest-leverage moves a program manager or PMO lead can make. It standardizes output across the team, accelerates onboarding, and gives you a single artifact to iterate on when governance requirements change.
How often should I update my prompt library?
Review it quarterly at minimum. Model behavior evolves as vendors release new versions, and your workflows evolve as your PMO matures. A quarterly review with 15 minutes per prompt keeps the library sharp without becoming a burden.
Can I use these prompts with Claude, ChatGPT, and Copilot?
Yes — all the prompts in the library below work across Claude, ChatGPT, Microsoft 365 Copilot, and Gemini. Model behavior varies slightly (Claude follows structured instructions more literally; ChatGPT has broader ecosystem integrations), but the prompt structure works everywhere.
Want to go deeper on AI-assisted delivery leadership?
Join Claude PM Pro — a 12-module program teaching senior PMs how to lead enterprise delivery in an AI-enabled environment.