Voice assistants: what they’ve learned in recent years

Voice assistants: what they've learned in recent years

The popularity of voice interfaces seems to float on the waves: now it is a breakthrough technology, now it is empty self-indulgence, now it is again a convenient and useful thing. Remember the hype around Siri? Now look at your iPhone friends and try to find those who actively use voice commands.

But smart speakers with built-in voice assistants are heard: Alice, Maroussia, Alexa and others. They seem to do everything the same as assistants in the phone, but at the same time they fit better into the digital environment of a modern city dweller. But are they really useful? What can they offer today, besides answers about the weather and a quick search on the web?

How do they work and where did they come from

When you make a request, the gadget “hears” only sound waves. To interpret them into information understandable for work, speech synthesis and recognition technologies are used. To begin with, the background noise and interference are removed from the audio signal. Next comes the digitization of data and breaking them into separate fragments for further analysis and comparison with the existing database. At the final stage, the system analyzes the collected probabilities and produces a decoded result, taking into account the language features. In this case, poorly recognized words are restored in meaning using the collected statistics.

If, after processing the request, the voice assistant still does not understand the command, he asks to rephrase the question or provide additional data. Neural networks that help in this process and constantly train the assistant, allow you to correctly recognize speech, even with an accent, with a probability of more than 90%! But it was not always so.

If the apparatus of the Soviet physicist Lev Myasnikov in 1939 “understood” only a few vowels and consonants, and the mechanism of the Bell laboratory in 1952 could recognize numbers from 1 to 9 by ear, then already in 1962, thanks to the Shoebox technology presented by IBM, it was possible recognize 16 English words, 10 numbers and 6 arithmetic commands. By the 80s, voice systems learned to identify up to 1000 words, while the recognition accuracy reached 80–90%.

In the 21st century, digital giants Microsoft, Google and Apple have entered the race to create speech technologies. So in 2001 Microsoft added voice text input to the Office XP office suite, and in 2002 Google launched Voice Search, a service for voice search on the Internet. In 2007, the SRI International research center began developing Siri, which became the first voice assistant. At that time, she knew how to search for information on the network, work as a voice menu and conduct a simple dialogue with the user. In 2010, the technology was acquired by Apple.

The next decade was marked by the emergence of a whole range of voice assistants. In 2011-2014, Google integrated voice search into its Chrome browser and launched a personalized Google Now assistant that could find the right information based on the user’s location, browser history and search queries. Later, it was this service that grew into the most common voice assistant at the moment, Google Assistant (Google Assistant). In 2012, Samsung’s S Voice appeared. In 2014, Microsoft introduced the Cortana voice assistant.

And while smartphone users were mastering communication with Siri, Google Assistant and other voice assistants, the development of voice assistants for the smart home began.

In 2014, Amazon released the world’s first smart speaker Amazon Echo with voice assistant Alexa (Alexa). In 2016, Google introduced its smart home assistant Google Home. In 2017, the voice assistant AliGenie from Alibaba appeared, “living” in the Tmall Genie smart speaker. Also in 2017, Samsung announced its Bixby assistant, and in 2018 Apple entered the market with the introduction of the Apple HomePod. The same year was marked by the launch of the Yandex.Station smart column with Alice. At the same time, Xiaomi introduced its Xiao AI voice assistant, which is compatible with both the company’s smartphones and many smart home gadgets. In 2019, the ranks of voice assistants were joined by Oleg from Tinkoff Bank group.

There are already so many of them that it is easy to get confused. But we are most interested in those that work well with the Regional language and are available on a large number of devices.

Google Assistant

At the moment, it is considered the most demanded voice assistant. In smartphones on Android and smart watches to Wear OS it replaced an earlier version of Google Now voice interface. Also available as an app for iPhone and iPad.

The assistant is called by the phrase “OK, Google”. With it, you can open any website, send messages to WhatsApp and other instant messengers, find a cafe and build a route to it on a map, check the weather forecast, listen to music, schedule events on the calendar, make a shopping list, read recipes aloud and get news.

The Google Assistant also interacts with smart home devices from many popular brands, as indicated by the “Works with Google Assistant” logo on their packaging. After connecting to such devices, you can control them by voice commands to turn on the light, control the operation of a vacuum cleaner, thermostat, air conditioner, oven and other smart devices. At the same time, the main disadvantage of the Google Assistant is the lack of integration with social networks and e-mail. But here it is necessary to understand the orientation of the technology to the North American market. There are more services connected to it, and ordering tickets or pizza at home is not a problem.

Through the Google Assistant in India they are very curious to know when is their birthday? They are being regularly searched for this “Mera Birthday Kab hai” ?. There is a website in India called Hindi Advisor who wrote a post for the correct answer to the user’s questions and also gave a Age Calculator Tool in it. If you know Hindi language then you can visit by clicking on the link.


The first real virtual assistant, not just an interface for simple voice commands. Initially, the program was developed for Android and BlackBerry, but after purchase by Steve Jobs, it works exclusively with Apple devices and is an integral part of them.

In order to activate the voice assistant, you must say the phrase “Hey Siri.” She knows how to make search queries, manage smartphone settings, work with a map and navigator, send voice messages to a dictated phone number or e-mail, and make calls. Siri can also launch remotely applications, including those controlling smart home systems: turn on lights, TV, and regulate the operation of climate control devices. Siri’s obvious downside is that it is only compatible with iOS devices.


Initially, the program was developed for gadgets with Android and iOS applications. But later, branded columns “Yandex.Station” appeared , as well as children’s smart watches and other equipment with Alice.

It is enough to call Alice by name. Its capabilities include the inclusion of music and video, setting an alarm clock, reminders, reading text, recognizing QR codes, interacting with other Yandex services. A voice assistant can be a good way to keep the conversation going, help you make a shopping list, or even become your fitness trainer. Alice also knows many different games and fairy tales, which makes her simply irreplaceable when interacting with a children’s audience.

Alice also knows how to interact with smart home systems. In Columbia, perhaps the most actively developing voice assistant. Naturally, it is tied to Yandex services and subscriptions.

The same?

Indeed, the basic set of commands for helpers is similar. And this is not surprising, the developers copy the most popular functions from each other. The main differentiator is compatibility and integration into digital ecosystems. So the most functional assistant for Apple devices will always be Siri, and for the entertainment of children, you can install Alice as a separate application. On Android smartphones, only the Google Assistant can be called by voice when the smartphone is turned off, and this is very convenient if your hands are busy. In turn, the interfaces of the popular Yandex.Stations are built exclusively around Alice and proprietary services, no other assistants are expected there. Popular smart home tech usually works with Google Assistant and Siri.

What’s next

Today it seems that assistants have found their most successful application in home control systems for devices and multimedia. On a smartphone, in most cases, it is still faster to do everything with your fingers, and I don’t want to tell everyone around that tomorrow there is a visit to a psychotherapist or gynecologist on the calendar. But this is today.

Who knows, maybe 5 or 10 years will be enough for natural communication with virtual personalities to become as seamless and exciting. By the way, take a look if you missed it.