Fable
Three gates before a new machine enters the factory
The textile mill originally used only one old machine. It was slow, but no one worried it would weave strange patterns.
Later, suppliers kept sending new machines—some faster, some cheaper, some able to weave complex patterns. The workshop director wanted to put every new machine onto the production line as soon as he saw it.
The first new machine did increase output, but on the night shift it wove a batch of expensive fabric to the wrong specifications. It performed well in the sample room, and only exposed the problem after entering the production line.
The factory manager first asked the director for a verbal guarantee. Of course the director guaranteed it, because he only looked at speed and did not bear the loss from returns.
Later the factory set up three gates: technical testing, risk review, and business approval. Every machine had to state its purpose, material sources, trial production results, limitations, and fallback plan.
Low-risk machines go through a fast track; high-risk machines require small-batch trial production and quality inspection sign-off. Every approval records the version, owner, and validity period. The workshop supervisor initially complained that it was slow, until he found that the approval log actually protected him. When problems arise, everyone can see who approved, on what basis, and how to roll back; when no problems arise, there is also evidence for confidently scaling up production.
Once a supplier upgraded a part; the machine's name did not change, but its behavior did. The approval process required re-evaluation and blocked a potential accident. Later the factory added another rule: approval is not a permanent pass. If the machine changes parts, the fabric changes source, or the use shifts from the sample room to customer orders, it must go through the gate again. An old signature can only prove usability under old conditions; it cannot vouch for a new scenario.
The workshop did not stop innovating because of this; instead, it now knew which machines could be confidently scaled up and which could only be tried in a sandbox. The plant manager said: bringing new machines into production should not rely on excitement and verbal promises, but on a reviewable approval chain. Later, the supervisor wrote this experience into the daily rules: first look at the scenario, then look at the boundaries, and finally look at who can take over if it fails. Only when all of these are clear will a beautiful demo turn into deliverable work. During reviews, they would also compare against old incidents to confirm that the new rules were not just good on paper, but could actually block the same mistakes on a chaotic shop floor.