Process improvement using Artificial Intelligence

[shukla7789]

Digital transformation is the observation of a process to be analyzed, on which we can identify a series of improvements to the process. These improvements can be with a combination of digital solutions, such as IoT (Internet of Things), Artificial Intelligence, etc.
Ruben Saavedra, Operations Manager PowerData Chile

Let's analyze, for example, a public transportation process using buses and analyze possible questions that we should ask ourselves to improve the service.

How do I improve the quality of my route planning?
How does information improve the person on a bus?
How do I identify fare evasion (non-payment) on a bus in real time?
How do I improve the quality of information provided to passengers?
How do I identify that a driver is not drunk or drugged when driving?
How do I respond to a person who is mute? And many other self employed database questions
An easy answer would be to put as many people to deliver these services, but clearly this solution does not scale, since many people would be needed to deliver the appropriate information and it would be very expensive. This is where these improvements can be made, for example, with the use of Artificial Intelligence (AI) algorithms , that is, developed applications that tend to do what a person would do, but through an application. Normally a person performs activities such as seeing, speaking, listening, etc., so what an Artificial Intelligence application does is try to simulate a human action (intelligence) but through a program (artificial), hence the name artificial intelligence.

Now let's apply AI to the fare-evasion improvement process mentioned above. To identify fare-evasion, I need to know how many people pay and how many people are on a bus at a given time. With artificial intelligence, you could place a camera at the bus entrance, so you can count how many people get on and how many people get off, if I can also know the position of a bus with GPS, and if I know that each person pays their ticket. This data, the time and route, are recorded, and then simply by correlating them (payment and counter), I can know what the fare-evasion is in real time.







You may be interested in reading:
The power of data-centric artificial intelligence







Based on the conclusions obtained from the data, I will be able to create public policies that better manage evasion. But not only that, if I manage to identify the age group and sex of the people, I will be able to know which sex evades the most, what age group evades the most, in which geographic location they evade the most, at what time, on what route, etc.

One of the important aspects to take into account in these solutions is to protect people's privacy. Who gave the right to record people? Clearly, there are questions that must be regulated here, but the same AI algorithms can hide the person's image, so they only know if they paid, did not pay, their sex or age.

Now, how can an algorithm detect how old I am and how much certainty there is in that conclusion? Well, AI algorithms only work with optimizations that mean a probabilistic result, so there is no certainty that the answer is correct. This also happens to us as human beings, hence the famous phrase ' to err is human'. For this reason, we must use it with caution, since it will not necessarily always identify the correct thing.