I recently read an article about a robot that can be controlled by human brain signals. The human, in this case a paralyzed individual, wears a skull cap with electrodes that picks up and amplifies the electronic signals emitted by the brain.
This is difficult, but fairly straight forward. The subject thinks about an action such as moving a finger and the scientists isolate the brain signals associated with the thought process. They write some code which translates the received signals into action commands for the robot. When the subject thinks about moving their finger the robot finger moves.
The challenge? Cutting through all the noise generated by our brain. When we think about walking we don’t focus on it consistently. We simply initiate the process and then leave it to some lower level automated process to keep it going. We think about it and then hop back and forth to and from other things.
Initially in these experiments on human thought controlled automation the human would be forced to keep sending a strong signal. To stay focused on one task consistently. That is hard to do and tiring for the human.
The breakthrough came when, instead of having the human focus on the task the whole time, they kick the task off and let the robot keep going until twww.logility.comhe human sends a stop command.
This impressed me because these same automation processes and rules apply to supply chain. The best systems continue to operate until a person is needed. The machine (a computer in this case) can execute a set of commands and manage the minor course adjustments within an acceptable predefined range and we (the human factor) are called in to action when a special event or exception is triggered.
This is the goal of exception management; to accelerate performance and focus valuable human resources. I want the plan to handle normal noise and I want to be alerted when something outside the norm happens.
The human is still very much required in planning and execution systems. We have found that when a system tries to automate all decision-making and become a ‘magic black box’ it almost inevitably fails. There is a balance between how much process to automate and how much intervention is required in every supply chain. The level and balance is different for each organization and business process.
How do you know whether an intervention point is ‘good’ or ‘bad’? It really comes down to value add. Is that intervention adding value? Is it reducing risk? Is it enhancing revenue or reducing cost?
Each year we get better at automating our supply chain processes. It is a virtuous cycle of available data, faster cycles and better science that marches ahead. Each time we take a stride we have to re-find the balance between what is valid, valuable process automation and what is not.
The modern supply chain is filled with uncertainty, risk and volatility. It is also filled with opportunity; opportunity for machines to do what humans are less efficient at. The trick is to define where we are less efficient and then program machines to send the right alerts, when we need them most, to allow us to investigate and add value, to move the business forward for success.