| Time | Destination | Status |
|---|---|---|
| 09:00 | Organisational AI adoption | Departs |
| 11:30 | AI governance | Departs |
| 14:00 | Pilot to production | Departs |
| --:-- | The pilot nobody owned | Cancelled |
| Time | Arrived | Status |
|---|---|---|
| --:-- | EU AI Act Compliance Guide: What Polish Companies Need to Do Before August 2026 | Arrived |
| --:-- | AI Agents in Business: A Strategic Perspective | Arrived |
| --:-- | AI Funding and Grants for Polish Companies in 2026: A Practical Overview | Arrived |
| --:-- | Common AI Implementation Mistakes: 7 Lessons | Arrived |
- 20
- Years engineering
- 100+
- Engineers led
- 3
- AI apps shipped
- 6σ
- Lean Six Sigma
Most AI pilots never arrive.
Not because the technology failed. The demo worked. The budget was approved. Somebody shipped something in six weeks and everybody in the room clapped.
Then the person who cared moved teams. Nobody owned it. There was no line for it in anyone's roadmap, no one on call when it broke, and no honest answer to what it cost to run. So it sat in a repository until people quietly stopped mentioning it.
That is an organisational failure, not a model failure. It is also the only part of this I find genuinely interesting — and the part almost nobody is selling you help with.
Departures board updated when something actually changes.
Who is telling you this
I have been an engineer for twenty years. Right now I lead AI adoption for more than a hundred engineers at a large insurer — the governance, the tooling decisions, the arguments about what we should not build.
In the evenings I build and ship AI applications of my own. I hold a Lean Six Sigma certification and I have run my own company, so I can read a P&L as well as a pull request.
I have no client logos to show you. I am early in this and I would rather say so than borrow somebody else's.