No single change broke the project. The total did. Every change request was small. Every one was reasonable. Two years and four hundred changes later, the project was unrecognisable and twelve months late. A study of incremental scope creep when change control exists only on paper.

One person held it together until the day she didn't. A project ran on the heroics of a single project manager for eighteen months. Then she left. We look at what was actually happening underneath the heroism, why nobody noticed in time, and the structural choices that made the team dependent on one human's stamina.

The risks that kill projects are usually the ones the team is too close to see. We explore how to use AI as a deliberately naive outsider, scanning your plans, contracts and tickets for the assumptions you've stopped questioning. With a workable prompt pattern you can run every sprint.

Most status reports are theatre, green, green, green, until the week before launch. We look at how AI can pull honest signals from commits, tickets, calendars and chat, and turn them into an update your sponsor can actually trust. Plus the politics of telling the truth faster than your organisation is ready for.

The same project update needs to land very differently with the CEO, the engineering lead, and the client's procurement team. AI makes that translation cheap, but only if you brief it well. We cover the audience profiles, the tone controls, and the one mistake that makes every version sound the same.

Can a daily AI summary replace the morning standup? In some teams, absolutely. In others, you'd be removing the only 15 minutes of human contact left. This episode is a practical decision tree: when to drop the meeting, when to keep it, and how to design an async version that surfaces blockers instead of hiding them.

AI can replace the standup as an information ritual, but not as a social ritual. The PM's job is to know which one their team actually needs, design the async version to surface blockers rather than hide them, and keep enough synchronous human contact to preserve cohesion.

Half your team is now pair-programming with Copilot, drafting docs with Claude, and running their own little agent workflows. So what's left for the project manager? This episode maps out the parts of the role that AI compresses, the parts it amplifies, and the new responsibilities nobody put in your job description.

AI is not removing the PM role, it is changing which parts of it matter. The work the PM did out of necessity is being compressed, the work that only the PM can do is being amplified, and new responsibilities, especially around quality control of AI output, are arriving without being added to anyone's job description.

Everyone left the kickoff feeling good. The objectives were broad, the timeline looked clean, and the stakeholders were aligned, or so it seemed. Six weeks later, three departments had three different ideas of what the project was supposed to deliver. This episode traces the confusion back to a kickoff meeting that covered the slides but avoided the real conversations. Listeners learn what a project charter should lock down before work begins, why assumptions need to be surfaced and documented, and what questions every kickoff meeting must answer.

Lesson: A kickoff meeting that does not surface the hard questions, assumptions, scope boundaries, and decision rights, creates the appearance of alignment without the substance. A real charter forces those conversations before the work begins, not after the cracks appear.

The Project Files is a storytelling series from Project Management Insights. Each episode follows a real project scenario where something goes wrong, gets complicated, or almost falls apart. The stories are fictional but the lessons are real, drawn from the challenges that show up on projects every day. If you manage projects, work on them, or lead the people who do, these stories are for you.