A personal assistant like Siri works thanks to the following technologies:

• Speech to text (STT) engine. The engine converts the user’s voice to text. The voice can be an audio file or a user’s speech stream.
• Text to speech (TTS) engine. Converts text to speech. That gets particularly useful when driving or cooking, so the user does not have to stop what he is doing to interact with the voice assistant. Also, it plays a big part in humanizing the assistant.
• Tagging (Intelligence). Tagging helps a voice assistant to understand what a user wants. For example, the user might ask, “Will I need an umbrella tonight?” Then the tagging engine can tag the information with weather or calendar info tag.
• Noise reduction engine. There is hardly ever a quiet and perfect environment for voice requests, there is always going to be a car moving by or a dog barking. So the noise reduction engine not only eliminates the white noise but also helps your assistant to understand you.
• Voice biometrics. It is a way of authentication, so your assistant could recognize your voice and respond only to your commands. Siri actually has it, you can teach how you say the words “Hey Siri”.
• Voice recognition. Machine learning component that drives all of the personal assistant mobile apps. This technology lets the assistant understand what you’re saying, basically it puts meaning behind your words.
• Speech compression engine. This engine comes particularly useful because it provides users with the fast output. It compresses the user’s voice so it gets sent to the server faster. You can use the G711 algorithm, that doesn’t lose the data, for this purpose.
• User interface. UI for voice assistants consists of two parts the voice and the call out. The voice part is what the user hears as a result of his question and the call outs are what he sees on the mobile screen.