AI prompts for project managers that turn status chaos into clear updates
Project managers spend hours turning scattered updates into plans, status reports, and stakeholder comms — exactly the synthesis work AI is good at, if you prompt it right. The difference between 'write a status update' and a usable one is structure and audience: who's reading, what decision they need to make, and what tone fits. These templates give the model that context so it drafts something you can send after a quick edit, not start over from.
Last updated · By the Prompt Orange team
Top prompts for project managers
1. Draft a project plan
“Make me a project plan”
Too vague—AI has to guess what you want
“Act as an experienced project manager. Create a phased project plan for migrating a 40-person company from Google Workspace to Microsoft 365 over 10 weeks. Break it into phases (discovery, pilot, migration, cutover, hypercare) with the key tasks, owners by role (IT, comms, dept leads), dependencies, and the top risk per phase. Present as a table. Flag the two milestones most likely to slip and why, and where I should build in buffer.”
Specific, clear, ready to use
2. Write a stakeholder status update
“Write a project status update”
Too vague—AI has to guess what you want
“Write a weekly status update for senior stakeholders on a website-replatforming project. Inputs: on track for the Aug launch; design sign-off complete; one risk — the payments integration is waiting on a third-party API key (owner: vendor, due Friday); budget at 48% spent, 50% through timeline. Structure: a one-line RAG status, key progress (3 bullets), the one risk with mitigation and owner, and what I need from them this week. Keep it scannable in 30 seconds and lead with the headline, not the detail.”
Specific, clear, ready to use
3. Build a risk register
“List some project risks”
Too vague—AI has to guess what you want
“Build a starter risk register for launching a new mobile app in 12 weeks with a team of six. Identify 8–10 realistic risks across scope, resourcing, technical, third-party, and external categories. For each: a clear risk description (cause → effect), likelihood (H/M/L), impact (H/M/L), a concrete mitigation, and an owner by role. Present as a table sorted by likelihood × impact. Don't list generic risks like 'things might go wrong' — make each one specific to this kind of project.”
Specific, clear, ready to use
4. Run a retrospective / lessons learned
“Help me run a retro”
Too vague—AI has to guess what you want
“Design a 60-minute remote retrospective for a project that shipped late but to good quality. Goal: surface honest lessons without blame. Give me a run sheet with timings: a psychological-safety opener, a structured activity to gather what went well / what didn't / what we'd change (with the exact prompts to put on screen), how to group and dot-vote themes, and how to turn the top three into owned action items. Add two facilitation tips for keeping a senior stakeholder from dominating.”
Specific, clear, ready to use
5. Summarise meeting notes into actions
“Summarise these notes”
Too vague—AI has to guess what you want
“Turn the raw meeting notes below into a clean summary for people who weren't there. Output three sections: Decisions made (with the rationale in a few words), Action items (as a table: action, owner, due date — infer owners where the notes make it obvious, flag where unclear), and Open questions still to resolve. Keep it factual, don't invent anything not in the notes, and put the most consequential decision first. Notes: [paste].”
Specific, clear, ready to use