Alexa

Alexa is a program from Amazon that lets you talk instead of typing. You say a sentence, the device converts it into text, looks for the matching action, and replies in an artificially generated voice. Alexa became well known through the speakers of the Echo line, which have been sold since 2014. But Alexa is also built into remote controls, cars, headphones, and TVs. Typical tasks include starting music, setting an alarm, reading out the weather, or switching on a lamp. The device only listens for a specific wake word, usually the name “Alexa” itself.

Why voice commands became a business model of their own

In 2014, Alexa was one of the first attempts to operate a computer without a screen or keyboard. Amazon sold the Echo speakers very cheaply, sometimes below production cost. The idea behind it: whoever orders by voice buys from Amazon. According to the company, over 500 million devices with Alexa have been shipped.

This idea only partly paid off. Most people use Alexa for timers, music, and weather, hardly for shopping. Reports from 2022 spoke of billions in losses in the responsible department. Alexa is therefore seen in the business press as an example of the fact that high user numbers don’t automatically mean money.

Alexa is also important for the debate about data protection. A microphone that constantly listens in on the living room raises questions. In 2023, Amazon had to pay around 25 million dollars in the US because voice recordings of children had been stored for too long. Cases like this still shape how critically voice assistants are viewed today.

From wake word to answer

A small detection process runs constantly on the device, only listening for the wake word. Only when it is triggered does the recording begin. The recorded sound is sent over the internet to Amazon’s data centers. There it is first converted into text, a process called speech recognition.

After that, the intent behind the sentence has to be recognized. “Make it brighter” and “Turn on the light” mean the same thing but sound different. The classic Alexa worked here with fixed programmed patterns and a list of add-on functions called skills. If a sentence didn’t fit any pattern, the standard reply came that Alexa unfortunately didn’t know that.

Since 2025, there is the revised version Alexa+. It relies on large language models, that is, AI systems that have learned from huge amounts of text how language is structured. Such models no longer need fixed patterns and also understand unusual phrasing. In return, they are more expensive to run and can make up things that aren’t true. In the end, the answer is converted back into speech and output through the speaker.

Alexa in the living room and in the headlines

In everyday life, most people encounter Alexa as a kitchen timer, radio alarm clock, or light switch. Anyone setting up a connected home links lamps, power outlets, or heating systems to the assistant. Cars and TVs often have the feature built in as well. The direct competitors are Google Assistant and Apple's Siri.

In the news, Alexa mainly comes up in three contexts: data protection proceedings, layoffs in Amazon’s devices division, and the race among corporations for AI assistants. A common misconception here is to equate Alexa with ChatGPT. Alexa is first and foremost a control system for devices and services, whereas a chatbot is designed for conversation and text. With Alexa+, however, the two worlds are clearly converging.