What is text-to-speech (TTS)?
Text-to-speech turns written text into spoken audio. It gives a phone system or voice agent a voice without anyone recording each sentence in advance.
Why it matters
Recorded prompts only cover what was scripted when they were made. Text-to-speech lets a system say things that change from call to call, such as a name, a balance or an appointment time, and lets an organisation update wording without booking a studio. The quality of the voice also affects trust: flat or mispronounced speech makes callers less willing to continue with an automated service.
How it works
The system first normalises the text, expanding numbers, dates and abbreviations into the words a person would say. It then works out pronunciation and where stress and pauses should fall. A voice model generates the audio waveform from that plan. For live calls the audio is produced and sent in small chunks, so the caller starts hearing the sentence before the whole of it has been generated.
Example
A clinic’s phone line needs to confirm an appointment. The system holds the text “Your appointment is on Tuesday 3 March at 2.15 pm with Dr Rao.” Text-to-speech expands the date and time into spoken words, pronounces the name using a stored spelling hint and plays the sentence to the caller in the same voice used for the rest of the call.
