Hidden waste inside the process

Every day we receive packages from Amazon and other delivery services. These days, even more. It happens more or less the same way - the delivery person comes, gives you the package, you sign for it, and then you enjoy it. Now imagine, if the day after another guy from Amazon would come to your house and ask you to confirm that you REALLY got the package. That would be weird and extremely inefficient for Amazon.

However, this is precisely what happens in all other industries when a work result should be verified. A contractor or a worker finishes a task and someone has come to a some place and physically check it before we proceed to another tasks. And this creates an inefficiency because usually the validator is a higher ranking manager and in an overload condition the system can stall due to a bottleneck - the time of validator. And also the validation happens in synchronised manner - therefore, you have idle time problem.

In a situation illustrated above the validator has to physically check the contractor work results - this happens in many industries. And the validator cannot be at two places at the same time. This means someone will have an idle time waiting for the validation to happen. In many cases, where the regulations require it, the validation should be done via physical presence, but does it need to be in all non-regulatory cases?

And again if we apply the same trick as in the previous post and introduce a different type of categorisation, then we can be looking at a different picture.

Now the premise is different - the validation tasks are split into two categories - the ones that require physical presence or the ones that don't (first 1-4 validations might require physical presence, but later it is not needed). Now, the contractor or any worker after finishing his task needs to capture it via video recording or a snapping a photo and submit as a result validation. The tasks that require physical presence will follow the normal route, but the tasks that don't can follow entirely different workflow. We achieve that by breaking down the validation work into smaller sub tasks and automating all possible routines.

In situation above the validation process has been split into several automated stages. First, the contractor submits the work results via mobile app (photo, video, 3D capture), then the results are pre-validated by an algorithm. The task of the system is the detection of anomalies and validation of the results via a combination of image recognition and decision systems. Eventually, the work result is classified depending on custom logic per each work, and depending on the result, the system can route either for on-site personnel to do manual work, to ask for a human to validate the content remotely by watching the video or directly passing the results as accepted.

This way we create a possibility to program validation behaviour and create parallel and scalable work validation system. Capturing work results in digital format will then allow us to get new insights and make better decisions.