Introducing Multimodal Conversational AI

Описание: We are pleased to announce a significant advancement for our Conversational AI platform: the introduction of text and voice multimodality.

Our AI agents can now seamlessly process both spoken words and typed text inputs simultaneously, leading to more natural, efficient, and resilient user interactions.
This development addresses common challenges in voice-only systems:

Enhanced Data Accuracy: Transcribing complex information like email addresses or order numbers via voice can be prone to errors. Multimodality allows users to type these details for perfect accuracy.

Improved User Experience: Inputting sensitive or lengthy data, such as credit card numbers, is often faster and more comfortable via text.

Greater Interaction Flow: Users can effortlessly switch between voice and text, choosing the most convenient input method for the context, making conversations smoother and more intuitive.

Our new multimodal capabilities are:

Easily Configurable: Enable text input directly in the widget configuration.
Widely Supported: Natively available through our SDKs, WebSocket, and the embeddable widget (requiring just a single line of HTML).
Versatile: Switch to a text-only mode for traditional chatbot experiences if desired.

This innovation builds upon our existing strengths, including best-in-class voices in over 32 languages, advanced speech-to-text and text-to-speech models, and robust deployment infrastructure with Twilio and SIP trunking support.