Generative AI has eclipsed technologies that make life easier through computers, such as services and processes. Third millennium solutions that serve users almost infallibly.
For those who use Windows, you know that by pressing CTRL+ALT+DEL you can see processes and services. Usually you manually close processes when the overlying program no longer responds to inputs…for example an Excel file that crashes…or RStudio / VScode that has been training a statistical model for a number of minutes that seems suspicious.
You may accidentally delete the process that shows the graphical interface on Windows. And panic can actually start, but it can be resolved by restarting the PC.
In general, a process means an algorithm that operates behind the scenes (background), in a non-visible way (e.g. a window)
Deleting a service, on the other hand, can make a PC practically unusable, requiring a system restore, therefore a more elaborate operation than a reboot.
On Windows, services act as servants, that is, they allow the user to execute, at machine level your requests. Network or connectivity requests, hardware, peripherals. To clarify, the mouse, a physical object, serves your movement request to the PC (or rather, to one of its peripherals) via a service: a kind of angel, i.e. a driver.
Drivers sometimes cause harm, but this is not why services are generically called demons, which according to Greek mythology had a neutral connotation, as they performed tasks that the gods did not want to deal with. An example of a hated driver? That of some printers. On Linux or other operating systems they generally cause more problems. One of the truest reasons for their limited diffusion as non server.
While with the process you usually interact via an interface (e.g. Excel, program with a graphical interface) and occasionally, depending on the activity, with the service you don’t. At most it requires an installation and then it has complete autonomy: it starts every time the PC starts. It only requires power and in some cases connectivity.
All this to say that IT services at the service of operations have existed for about half a century and this must be remembered when we talk about generative artificial intelligence, because lately we talk about AI assistants and AI agents, even in a confusing way.
The assistants are used when you make a request to them, via some type of terminal or graphical interface, the agents operate autonomously following, say, a first configuration or start-up. In other words, assistants have reactivity, agents have pro-activity.
Doesn’t it remind you of a difference between process and service?
As per the previous article, you can easily cite an example of an AI assistant: chatGPT. Much more difficult to cite, in February 2025, an example of an AI Agent that can be understood as quickly as chatGPT.
It could therefore be said that AI agents are a glorified service, as they even aspire to replace the workforce. Many companies are pushing hard on the fact that to have a very profitable and useful service for the company, it is necessary to insert generative AI, a discussion already addressed.
If in the mythology of ancient Greece demons served the gods, today, in a bizarre reversal of the situation worthy of psychometric studies, AI agents (venerated services) are seen as species of gods, culminating in the so-called Californian cult of artificial general intelligence (AGI). Literally a deus ex machina.
Attention: on the YouTube channel when I talk about statistical services, I don’t do it with the computer meaning of this article.
Do you want to understand which servant of the third millennium is best for you? Let’s hear from you in a free first call.