“Hey Siri, wake me up at 7:15 tomorrow”. Here we go! It’s been in every decent sci-fi movie beginning from Star Trek to Iron Man or Interstellar, the virtual assistant, who works as our digital PA and manages all of our dreadful everyday tasks. Not long ago it was only a dream but thanks to the newly established market with “machine learning” it made this dream come true. Let’s have a closer look.

Robots Galore

Our world we live in is full of robots. Period! Coronavirus topped it up and hence the year 2021 could quite easily mark the robot era. The main turnover thanks to the world pandemic however happened throughout the year 2020 but the consequences will be felt long after the virus diminishes. One of the fields that got a kick start was the one of virtual assistants together with areas of teleconferences, internet banking or gaming, but lets talk about the other fields another time and keep our focus on the Virtual Assistants. Almost every person has a virtual assistant (for iPhone users it is even part of the default setting) and its main task is to save time and money. It is tightly connected with 4th industrial revolutionthat has already begun. We hit the ceiling for human potential, more precisely Japan hit the ceiling. In the land of the rising sun, they can get as much out of you as 21% effectiveness. Unfortunately, during the 24-hour cycle, people have to sleep, eat, rest and kill their time on useless admin tasks that, at the end of the day, doesn’t bring any added value. If we want to move forward and push our boundaries, we have to outsource part of our everyday tasks to machines, in this case we are talking about automated virtual assistants. Don’t be fooled, all of the VA are not at all on par. Let’s explore for example J.A.R.V.I.S. (Just A Rather Very Intelligent System)

We’ll start from scratch; I like to use some real anchor points to visualise it better before moving to abstract levels. Virtual assistant isn’t anything else but a program running on a general computer. Well, that computer isn’t that general, quite often it is a supercomputer (high level of performance) and to call a virtual assistant “just a program” is probably a bit of euphemism also, but you get my point. Why do we need supercomputer then? Because virtual assistant has to do lot of calculations to reach the desired outcome. It’s making absurd amount of tiny little calculations that when put together it puts the CPU in high usage. And why does the virtual assistant have to do so many calculations? Very good question. It is trying to simulate human brain, well maybe not the human one because that is well above any computer’s wildest dream, so maybe lets say a fish brain, but the principle is the same.


The inner implementation can vary significantly in different solutions but the most common are backpropagation neural nets that enable to simulate “learning” process that move the whole neural net forward. When you write simple algorithm and run it, it behaves the same even when you run it 100x times in a row. We like this kind of determinism about the computer science, it allows us to be 100% in charge and control. On the other hand, it means that the algorithm never makes any decisions on its own and if it does make a decision, it will always be the same outcome. In the role of virtual assistant, that type of pattern isn’t very useful if the answer is for example always NO. Neural nets or more precisely the machine learning has the ability to learn with every new decision whether it was the right one or rather not. I’m not going to bother you with too many technical details but it’s no rocket science, the theory foundation was laid sometimes in the second half of the last century.

Thanks to the fact that the program of virtual assistant has the ability to learn, it can manage basic tasks. At the beginning its “mind” is very simple, just like a three months old baby you wouldn’t ask to create a slot in a calendar for a meeting with your friend at your favourite golf course. Over time you can teach it what the calendar is, what a meeting is, who your friends are etc. and it can help you with certain aspects of life that would otherwise have to be dealt with by yourself. It could set an alarm by itself for you, you can use a dictation feature to say a message and it will then translate it in to text and send it to your wife; it can even vacuum your floor at home (providing your robot vacuum cleaner is capable to be operated remotely). All this is truly happening and virtual assistants can complete some human tasks that are tedious and hold us back from more creative and satisfying pursuits.

Virtual bankers

Of course, it brings lot of drawbacks and the outcome isn’t always as peachy as we would wish for, but it works, people are using it and they even pay for it. Hard to believe? Try to tell “Alexa, play some music” and few developers just had their lunch paid for. Even in the Czech Republic there are not only teams but whole companies that make their living from creating and maintaining virtual assistants. Our job at the moment focuses on virtual assistant for Ceska Sporitelna and despite the fact that it’s nowhere near advanced as Siri, it’s great and very interesting work that pays the rent.

It is said “future is now” and if you are interested in this topic, just google IBM Watson Wingbot.AI, Mamma.AI, Alexa or even Tesla and its autonomous systems. By the way, Skoda Auto can now drive your car for you on the freeway. Isn’t that wicked? AI it’s here with us and it will never go away. Roll up your sleeves, because it’s all just beginning.

0 - 0

Thank You For Your Vote!

Sorry You have Already Voted!


Read more

Inverse Conway Maneuver!

“Any organization that designs a system (defined broadly) will produce a design whose structure is