Twenty years ago today, on August 29, 2006, I published the first posts on LeadingAnswers.com. That sentence makes me feel old. I started the blog to write about leadership and agile project management, share ideas I was experimenting with, and hopefully connect with other people interested in the same things. At the time, agile was still sufficiently new that explaining it to traditional project managers was a large part of the job. My first real article was called “What has Leadership got to do with Agile Project Management?” My argument was that effective agile project management was much closer to leadership than traditional project management. It emphasized people over tasks, empowerment over control, effectiveness over efficiency, and direction over speed.

Today we explore agentic pipelines, loops and recursive self improvement (RSI) via our old friends, PMP Questions. This is article 4 in my AI agents series. Start here for an introduction, see here for an explanation of what constitutes a real agent, and here for a discussion of how I use them and running AI locally. This article gets into how to plan and manage agent work; that’s what “agent orchestration” means. As a project manager, I suspect you already have the thinking skills required. You already know why one person can’t do everything on a project. You staff a project with specialists, engage each of them in the correct piece of scope, and put reviews between the stages that