I have recently been in contact with a Chinese factory to improve some of their processes. I visited their factory, spent a day learning about what they do, how they do it, what are some of their main challenges and how they thought new software tools could help.

Out of the 3 main aspects discussed on that day, only one seemed a real fit to involved development : building a new tool to replace a spreadsheet they currently use to manage and plan their daily manufacturing production. It is currently a painful manual process that could be automated and improved upon and later optimized.

The main one I want to expand on here is the 3rd project.

The problem:

Mistakes are made during the manufacturing process by workers following technical drawings.

This happens when the workers are not being careful and assume some of the settings (for example when differences are minors between to items, size material, ...)

The issue then leads to full batches of components that need to be discarded and redone.

The current process in that factory is:

1. A commitee reviews new technical drawings and make notes of the areas where special attention is needed.
2. When production should start for a specific piece, the worker receives a paper version of the drawing along with the notes made.
3. Production of the component
4. Quality checks on the pieces produced

That's possibly a good process, but there is still the issue coming up.

The discussion covered multiple aspects, mainly in uncovering the areas of special attentions, using AI or others, but ... those addressed a different problem, the one the committee is solving : identifications of point of interests.

My recommendation:

Do not think of any software solution.

While a tool solving the identification part could be useful, it would not fix the main underlying issue :

Details are overlooked, there is a missing step in the process to ensure nothing gets ignored, and no software could solve that (especially in this context where it is paper-driven).

If that is the main pressing issue, the process itself needs to be fixed. Should the notes be explained more explicitly , clearly, in a different way, ... when the worker receive the drawings? Should the worker be signing of on the list of details to pay attention to? ... Those are more direct and impactful changes to bring in.

Once an impactful and successful fix is found, it may be time to bring in software to help with the implementation or optimization.

Conclusion

Remember to consider what is the problem you are trying to solve and what is the root cause. Would any software really fix it?

If not,

Fix the issue first, then you can consider whether adding software is the solution.