AI for Project Management Capstone Exercise — AI for Project Managers

What you'll learn

Apply skills from across this course in a single realistic professional scenario
Produce a concrete, role-relevant deliverable using AI tools
Self-assess your output against professional quality criteria

Project initiation is where poor AI use causes the most downstream problems. Artifacts produced too quickly without PM judgment baked in become the wrong plan delivered with false confidence. This capstone tests your ability to use AI to accelerate initiation at pace without sacrificing the professional judgment that makes the artifacts trustworthy.

You are a project manager kicking off a six-month system integration project for a financial services client. The kickoff is in 72 hours. Four artifacts are expected by tomorrow afternoon. You will use AI to produce first drafts of all four, then mark exactly where the AI cannot be trusted to fill in what only you know.

Capstone Exercise

Accelerate Project Initiation Artifacts Under Deadline Pressure

Context

You are the project manager responsible for a six-month integration project connecting a financial services firm's core banking platform to a new third-party payments processor. The client is regulated, the integration touches live transaction data, and the governance structure requires board-level sign-off at the end of sprint 1. The kickoff is in 72 hours. The sponsor has asked for the project initiation document, stakeholder map, RAID log, and sprint 1 plan by tomorrow afternoon, and has made clear that incomplete artifacts will delay the kickoff.

Your Task

Use Claude or ChatGPT to first-draft all four of the following artifacts: (1) A project initiation document covering scope, objectives, constraints, and governance structure for this integration project; (2) A stakeholder map with eight named stakeholders classified by influence and interest level, with one sentence on the engagement approach for each; (3) A RAID log with five initial entries per category: risks, assumptions, issues, and dependencies; (4) A sprint 1 plan with ten tasks, owner roles, and estimated effort. Produce all four artifacts, then add one PM judgment note to each, identifying one content element the AI has generated that you would need to replace with project-specific knowledge before the artifact is ready to present to the client.

Your notes (optional)

Deliverable

Four project initiation artifacts: PID, stakeholder map, RAID log, and sprint 1 plan, each with a one-sentence PM judgment note identifying the content element requiring project-specific knowledge. Demonstrates the artifact-acceleration pattern from the course applied to a realistic, deadline-pressured initiation scenario.

Quick check

Across the four initiation artefacts, which element does this capstone single out as the least reliable part of an AI draft?

Select one answer.

Key takeaways

AI can produce first drafts of all standard project initiation artifacts in a fraction of the time manual production takes, but the PM's judgment is what makes those artifacts trustworthy: replacing generic content with project-specific context is the professional work that AI accelerates, not eliminates.
The RAID log dependencies category is consistently where AI-generated content is least reliable in integration projects: sequencing and technical dependencies require domain knowledge that a general-purpose model cannot supply from the prompt alone.
Deadline pressure is one of the highest-value contexts for AI-assisted PM work, provided the PM applies the verification discipline practiced in this capstone: drafting fast with AI, then annotating and correcting before the artifact reaches the client.
The PM judgment note discipline trains the habit of treating AI output as a starting draft requiring professional review, not a finished artifact requiring only formatting before delivery.