I tried that in a discussion. Several people read the same restart line differently. So yes, we've handed over an instruction that doesn't agree with itself.

Use the conflicting interpretations to revise the instruction into condition, permitted action and escalation criteria. After the responsible integrator reviews it, check understanding again through an unprompted teach-back.

Because I wrote the ambiguous instruction, I'm concerned that teach-back could feel like blaming the operator. How can I frame it as a check of the handover?

@SarahBell0676 For my own handover, I framed the discussion around locating gaps in the guide I had written. Asking where it required a guess helped operators contribute corrections without defending their interpretation.

Once the recovery sheet is clear and available at the controls, the team should no longer need to call you about these interruptions.

Not if the system can't tell them what's physically happening. Better wording can't create a recovery path for an unknown held-workpiece state.

You're right; I promised too much from a sheet. The guide should support defined recoveries and clearly escalate unknown states. Some calls are the correct outcome.

Our original handover only showed an empty gripper at startup. It never covered an interruption with the workpiece still held. That's a real gap.

Couldn't the operator just select the pocket they were on? Or is that exactly the guessing you're trying to remove?

Knowing an index doesn't establish whether the workpiece is held or the destination occupied. Recovery needs a defined way to determine those conditions and an integrator-reviewed action for the result.

@RobinChan1121 On my own cell, changing the index became folklore: ask the person who 'knows the trick'. Replacing that with a defined escalation made the uncertainty visible.

An operator also spotted that our screen and printed guide use different names for the same step. I wrote both. Lovely bit of self-sabotage.

Align the printed step names with the installed display and verify that any illustrations correspond to that version. This removes an avoidable translation task during recovery.

@RobinChan1121 Check the meaning too. Two identical labels won't help if they describe different physical conditions in different places.

Would you practise this on the machine, or talk through pictures first? I wouldn't fancy being surprised by an interruption in training.

You can establish understanding through offline examples before equipment practice. For the practical stage, use an agreed supervised procedure with defined conditions rather than unexpected fault creation.

What should I note during the discussion? Just whether they picked the right action, or where they hesitated too?