How does this work?

Voice Assistant as a Service

The Slang platform is unique in the way it enables adding Voice Assistants to your app. Just like AWS enables adding new devices to your "lab" by making use of pre-built VMs, the Slang platform includes a bunch of pre-built domain-specific Voice Assistants, that have been pre-trained, and provides a simple way to consume these Assistants inside apps belonging to those domains. The platform comes with

  • Pre-built domain-specific Voice Assistants that have been pre-trained and pre-configured to support a pre-defined set of user journeys and an extensive set of app states inside each user journey. Every Assistant comes with the following -

  1. Abstracts out the underlying deep tech components like ASR, NLP, and TTS so that the consumers of the Assistant don't have to think in those low-level terms.

  2. Pre-trained to understand multiple domain-specific user journeys and is also trained with relevant data from that domain (which can be optionally extended if required)

  3. Multi-lingual support out of the box for all the supported user journeys

  4. Every user journey comes with its own rich set of app states and app state conditions

  5. And every app state and app state condition comes with its own pre-baked conversational design constructs like

    • Prompts, which are questions that the app needs to ask the end-user

    • Statements, which are things that the app wants to inform the user

  6. Rich UI components that can be embedded inside your app, that is needed to interact with the end-user. It also includes sub-components to help the user get onboarded successfully and also helps them to get trained on what to speak and when to speak

  • Configuration options to customize the Assistants to suit the app's needs

  • Mobile and Web SDKs to allow the app and the Assistant to work together seamlessly

  • Auth management

  • Accounting

The Slang In-App Voice Assistants require an internet connection to work. It talks to a highly scalable backend that is optimized to provide sub-second responses to most user queries (depending on the network bandwidth)

From a developer perspective, there are two main components involved -

  • The Slang Console - where you select, configure and manage the Assistant and also your account details

  • The Assistant SDK - which is integrated into the app and exposes the programming interfaces for the app to interact with the Assistant and vice versa.

Let's now try to dive deeper into the most important concept from a developer perspective - User Journeys and App States