There are millions of chatbots and voice bots (conversational assistants) in the world, hosted on mobile phones, smart speakers and websites. They are hosted by many kinds of organizations - including governments, businesses, and non-profits. Today, each conversational assistant is independent of the others, even within an organization, and each one has its own expertise. What if they could collaborate, working together and sharing information? This would reduce implementation complexity, duplication of effort, and friction for users.

The Open Voice Network Interoperability Initiative of the LFAI and Data Foundation is working on enabling just these capabilities. Its goal is to make it possible for different voice and conversational AI systems, based on LLM's or other NLP technologies, to work together through the use of a standard open-source set of easy-to-use messages.The messages enable any conversational assistant that follows the standards to freely interoperate with other standards-using assistants to connect, communicate, and transfer content and control across assistants, platforms, and language models. Please join us for a workshop that will show developers how to use our specifications to create ecosystems of collaborating assistants, including both new assistants as well as legacy assistants.

This workshop will describe the messaging protocol and then introduce the Open Voice Interoperability Initiative implementations of the protocol. We start by reviewing the open source Interoperability Sandbox implementation. The Sandbox includes a web interface client as well as an assortment of interoperable assistants available for experimentation. We will also introduce BeaconForge, an open source assistant server, that can be modified to build interoperable assistants based on LLM models. Attendees will get a chance to build their own assistant during the workshop.