Preparing the published article.
Lessons From Failed Robot Projects
Most failures are predictable, and none of them are technical mysteries.
The recurring causes
- Part presentation was less stable than the demo assumed.
- Cycle time was promised without counting handshakes and settling.
- Nobody owned the cell after commissioning.
- The scope grew without a change log, and the original payback disappeared.
- Safety was assessed after the layout was frozen.
| Warning sign | What to do at the next gate |
|---|---|
| "We will fix it on site" | Insist on factory acceptance with real parts |
| No named operator | Stop and assign ownership before delivery |
| Cycle figures without basis | Demand the calculation and the assumptions |
| Changes discussed verbally | Convert to written change requests |
A failed project is rarely a bad robot. It is almost always an unmanaged assumption.