Miguel Marengo Canales

blog/Silodisa AI log

weekly log · issue 6 · the forklift with a failure

149 of 426 shifts: the forklift that worked with its seatbelt marked “no”

The Guardian cross-checks every forklift’s checklist against that shift’s pallet moves: 149 of 426 shifts worked with a critical failure reported. We don’t punish reporting; we punish working that way.

Miguel Marengo Canales
 |6 min |comments
start-of-shift checklist brakesyes parking brakeyes emergency stopyes seatbeltno extinguisheryes window · max 12 h pallet moves shift with failure critical "no" + moved pallets = counts
One shift, start to finishIllustrative. It counts only if one of the five critical lines says "no" and the truck moved pallets within the shift window.
in 30 seconds
  • The Guardian reads every forklift’s checklist and cross-checks it against the pallet moves of that same shift.
  • We don’t grade what gets reported; we grade working with the failure. Reporting is good.
  • From August 28 to September 26, 149 of 426 shifts worked with a critical failure reported, almost always the seatbelt. We’re still in the red here.
  • A single shift like that turns the alarm on, and with the alarm on the warehouse score can’t come out green.
149 / 426shifts with a critical failure, 30 days
5,538moves in shifts with a failure, 7 days
6 dayslongest streak
5critical checklist lines

Every morning, the forklift operator fills out a checklist: brakes, parking brake, emergency stop, seatbelt, extinguisher. “Yes” or “no.” Almost nobody reads it. We asked an agent, the Guardian, to actually read it.

Its first question was simple: if someone wrote “no” for the seatbelt, did that forklift stay parked or go out to work? I didn’t like the answer.

What counts and what doesn’t

A critical failure is a “no” on brakes, parking brake, emergency stop, seatbelt or extinguisher. A shift runs from checklist to close, 12 hours at most. And the shift worked if the same operator moved pallets in that window.

The rule we debated most: we don’t grade what gets reported; we grade working with the failure. A “no” with the truck parked is the right behavior. If we punished reporting, within a month every checklist would say “yes.”

moveddid not move countswith failure shown,not counted no failure did not work criticalmarked "no" all "yes" reverse alarm "no": shown, not critical
What counts and what doesn’tThe test cases written by hand before the code. Reporting is never punished; working with the failure is.

The numbers

From August 28 to September 26 there were 426 shifts that moved pallets. 149 worked with a critical failure reported. In the last week of that period it was 42 of 106, with 5,538 pallet moves. The most common failure is the seatbelt, and we’re still in the red.

30 days · Aug 28 – Sep 26 149 of 426 shifts 7 days · Sep 20 – 26 42 of 106 · 5,538 moves
Shifts that worked with a critical failureReal 2026 data. Full bar: shifts that moved pallets. Blue: those that did it with a critical failure reported.

Two trucks, A and B, reported a failure 6 days in a row. The checklist worked six times: someone saw it and wrote it down. What didn’t work was what came after.

streak: same failure truck Aseatbelt no no no no no no ok truck Bseatbelt + brake no no no no no no ok 6 days in a row
The streakThe 6-day streak is real; the position of the days is illustrative. Counted per truck and per failure, never per person.

An alarm you can’t average away

There are six graded numbers and three alarms that don’t add points. A single shift that worked with a critical failure turns the alarm on, and with it on the score can’t come out green. If the system couldn’t read the checklist, it says “no data,” not “off.”

6 graded numbers alarm: failure score can't be green 1 shift is enough · not averaged
An alarm you can’t average awayIllustrative. A single shift that worked with a critical failure turns the alarm on; with it on, the score can’t come out green.

The Guardian shows only the truck’s fleet number, never names. The problem isn’t an operator: it’s a process that lets a truck with a reported failure go out to work.

What we don’t know

  • The checklist is what the operator wrote; some “no” may be a typo-level mistake. Random floor checks come next.
  • We don’t yet measure how long a failure takes to fix: that number shows gray.
  • The 12-hour window is an assumption, written down with its example.

What comes next

The Guardian doesn’t stop forklifts: no agent writes to the warehouse system on its own. It puts the alarm where it can’t be ignored, and its owner decides. The real question is operational: do we let a truck work with the seatbelt or the parking brake marked “no”? BIEN y a Tiempo (done right and on time) starts with “right.”

In your operation, what happens when the checklist says “no”? Write to me.

Miguel Marengo Canales · Silodisa

weekly loglogisticsGuardiánsafety

Comments

comments · I answer every one. If you disagree, even better: say it with numbers.

  • loading…