10 golden tips for an AI Manager - how to manage AI projects

In this episode we dive straight into the everyday reality of running AI projects, with Dagmara Sysuła, Delivery Manager at Office Samurai, as your guide. If you’ve ever tried to run an AI project the same way you run RPA or IT delivery, you already know how that story ends. AI refuses to behave. And Waterfall? It taps out early.Dagmara walks through why AI needs a different kind of steering. These projects live on experiments. They shift. They surprise you. And unless you mix Agile discovery with the reporting discipline your sponsors demand, you end up with a confused team, an annoyed CFO, and a solution that never leaves dev.The four types of AI projects we keep seeing are:POCs and quick tests: two or three weeks to see whether the idea breathes or just gets sucked into the black hole.Full deployments: big, messy, political, and the ones that actually move the needle.Buying tools: buying the license is easy. Buying the knowledge to use it well is harder, and most companies skip that part.Strategy work: setting guardrails, deciding where AI fits, and lining up the ethics and processes. This shouldn’t wait until “later.” It has to run next to the early delivery work so the strategy is shaped by reality, not PowerPoint.That is why Dagmara came up with the 10 rules that can help put the puzzles of the AI project into the perfect place. Straight from the field, far from theory. We won’t spoil them that much in the description, so that you can get to know them yourself. ;) AI moves fast enough that these rules may look different a year from now. But for now, they’ll keep your AI projects from drifting into the usual swamp of confusion, hype, and half-finished work.Listen to the flaming discussion between Andrzej and Dagmara to deliver you AI projects better. And faster.