On the (un)importance of dimming light bulbs

Automation in the digital sphere becomes less about some one-time fix of a routine and more like a rebuilding of the IT process environment with a business.

Typically, the IT infrastructure and the user journey focus on what business thinks is important. It is trying to hide the management from the employees, and, because the goal of automation is far from an announcement, everyone manages the same way in digital. The business has all the information, but it may even not know what is needed for work, what information will induce more productive work, because the viable information is dumbed down by the equality of the decisions, all of those decisions for which a good manager finds himself walking to someone, asking and getting some mixed piece of information, partly full, partly operative, partly something else.

The business process is in the hands of the people, and all the stuff that is out of the scope is executable in different ways. When you do a system where everything is on display and measurable then you can debate on management styles, about how the process should flow. And this can be done after the manager has been administrating it to some level where the crucial management parameters showed up. It is better if you allow working effectively in some range — not a deadline, but a range because a process is a system graph where you can do a lot of one sub-process instances quickly, but the other thing would suffer badly from not being paid attention to.

So, it would be a system which helps managers avoid the deadlines by optimizing the execution time much more when it is needed because it does make sense to make some job in a deadline.

We may have some strict deadlines: for example, food expiration date. We may have some deadlines only because we must move the cargo package into the production site before another cargo arrives: we do not know when it will come but this can be calculated upfront, having the amount of cargo before and the required amount after. You do not know the sizes of upcoming orders, but you may take average orders from most frequent contractors and virtually place them, and voila — you have a prediction of the order, and you know that you will need this place right for the next batch. Like, for the second order, which will come in approximately three days — a soft deadline that can be manageable. If you perform everything along with such style of upfront analytics, you can handle the things with soft deadlines sorted by deadline time and by priority.

It all ends up that an employee work is becoming more like making a detail at a factory: he does something for a period, then he is said to do another thing or goes back to doing the previous thing — mid-level decisions are excluded from his process, and so is learning. And so is your learning curve as a manager, which makes up your ability to set up and maintain an educational process. Both managers and employees often do not want to do this.

How to help people learn from the system?

Well, by giving in the information. Given the information and an organized process, the person will gain knowledge of the process.

And because it is well-built, you will not think about traditional communication and coordination problems which will not arise — because they just do not arise, you know, in a well-built process where employees are constantly learning themselves.

But then all the other problems may arise, like the problem of system usability. Also, the employee may leave the job with the knowledge of that problem, trying to handle similar problems, and maybe not being able to handle them, because addressing those problems from the bottom level is hard.

So, when doing a project, one might think: maybe I am enhancing technology within an information system that is not good enough initially, and I am trying to cap this thing by performing almost impossible technical integration?

I mean, it is all possible with very pricey software that will manage everything, handle commands in a natural language and do all the yoga stuff for you while you are lying on a couch. You can execute it in the numerous fancy ways or just in a procedural way.

If you are stuck with the process, maybe it is not well-written, and there is a technology you already know subconsciously — you may not just need to “unload the truck” but “take a box” and “move it to the ground” and ultimately act like it is to be done by a robot.

Cargo movements can be done in such a way but not everything and every decision can be given to a robot or digitalized within a system where everything is optimized in this fashion. Maybe there are some curved surfaces to be worked around — some leather material work in the car interior can be done by people. Anyway, there are ways of working with leather so that leather does not tear up or lose its consumer qualities in any way, and these ways are suitable for machine operation. When it is done, management of those materials is built into the machine and you just drag and drop the materials on a product part, it looks them up at the market, helps to find solutions for switching the materials to another ones if it is a feasible option.

So, it manages all the stuff, its function is changing all the time and eventually, it is not just doing the job — the function of the software will be optimizing the process economically. It will function by maximizing output and minimizing input, not just processing the input in a predefined way.

We cannot predict where it will lead us but anyways it might be used in a soft way, where it just enhances the processes. But it is just too expensive to implement and maybe one should think of an old-fashioned, “human” way of optimizing the things logically, but with gradual digitalizing this same thought process in parallel.

You cannot be a stupid solution architect — if you do not know from the ground up how to make a product, you will not be able to produce it with a stable desired quality output.

You cannot programmatically reach success — you just can redefine it or move towards it.

So, success is in knowing what your success is, what is your potential, seeing the potential of things that can help you release your unique potential.

Face this idea:

In the procedures and specifics, you can submit some production parameters into the neural network, and the system will tell you that you have something that correlates with something. When is it necessary? Not when you just see with your eyes that the light bulb needs to be moved to the desktop, and it will immediately increase the efficiency of work — this is obvious. The production factor should be such that we do not see it with our eyes, because it may not be specific — for example, a general production light bulb has changed the lighting temperature, or something else. But to analyze that, you need to be changing it for a while, this light temperature, okay? Otherwise, when you give the neural network some unchanged parameters, they remain the same through time, because there is no data on the impact of changes, and it does not try to change them. The dynamics, of course, also need to be tracked with subtle parameters to be found.

In other words, the question is often voiced in a way that the effectiveness of the response depends on how we process the order, how we will respond on the phone, at what time we respond. No. The faster the response, the better. We’re building a just-in-time on-demand business and that’s it, guys. You do not need this — “at what time the operator responded” or this — “the data scientists found out the probability”. The correct thing is: your server was not working, you just had a bad service, and at this moment there were active competitors. Here are purely logical reasons that are solved by abstraction — initially, when creating a business, and further on. Abstracting is good, it is like when you are a scientist who has found an island in the ocean just by sitting down and analyzing the ocean currents. And the dynamic parameters go after.

I am not talking about the fact that dynamic parameters arise because of speed, due to lack of necessity to wait. That is, if you have everything simplified, you have a high speed, and experts are the ones who can simplify and analyze the problem to the level of simple things. Any problems can be decomposed to the level of simple things. If you have specialists, they will help you do such simple things. If you want to implement even a simple thing, simple thinking, in a large system — it will be innovation, it will be innovative thinking, in fact, without technology. In fact, when you have simplified everything to the level of an object, a product, an object of management — this will be innovative thinking. When you introduced the concept of a management object and gave it to the manager to actually start managing. Where there is management, there is responsibility. And if he does not manage anything, then he is not responsible for anything. Give the loader an understanding of management, he will put the goods more carefully because this is related to his responsibility.

Although the need for digitization is clear, it existed even before computers.