Cognigy's Technology Partner CoCo built a custom module for Cognigy.AI to help users easily connect and add conversational components, therefore simplifying the chatbot-building process from start to finish.
Building a good chatbot is not easy as many publicized failures show. Bot designers struggle to manage the many ways even a simple conversation can go. One key problem is that most designers start from scratch with every chatbot and there is only a limited amount of time in a project to get everything right. Cognigy and CoCo want to make bot-building even more accessible by simplifying the process from start to finish.
Who is CoCo? The Building Blocks of Conversation
CoCo may be a new name in the world of Conversational AI, but it’s built on founder Yaki Dunietz’s 20+ years of experience in creating and innovating conversational artificial intelligence. CoCos' goal is to change the field of conversational bots by unifying their design and sharing code, content, and technical advances.
Starting with pre-built blocks of dialog, or “Conversational Components” (CoCos), developers and conversational designers are freed from building generic conversations that fill the most basic parameters. CoCos specialize in providing basic building blocks and common conversation tasks like asking for a users' name, address, phone number or scheduling an appointment, allowing the designer to concentrate on the unique conversational aspects related to their business.
The API is designed to make it easy for developers to switch between the bot they’re building and an external, ready made one of CoCo's components. Generally, there needs to be a signal that determines who should answer a particular input, the calling bot or the CoCo component.
Conversational Components employ three simple signals that govern the flow: component_done, if the component has achieved its goal, component_failed, if the component could not achieve its goal, and out_of_context, if the bot should respond to the input for only the following turn. Once the relevant CoCos are selected, it’s time to assemble them into a cohesive bot: the task of piecing the components together. That’s where Cognigy comes in!
Cognigy.AI: An intuitive Interface with Advanced Coding Customization Tools
Because Cognigy’s platform is designed for building a conversational bot from start to finish, it is the perfect orchestrator. Together, they produce a multi-turn, context-sensitive conversation. Cognigy.AI has simplified the process that is usually involved in building a conversational flow. The intuitive interface makes it fast and frictionless when visualizing how a conversation will unfold. Nonetheless bot builders still need an understanding of the logic behind the conversation. When it comes to chatbots this includes elements like anticipating a variety of different inputs and designing the bot responses based on those. While many other systems require conversation designers to simply imagine the logic they build into the bot, Cognigy.AI makes it possible to visualize the flow itself.
When combining Cognigy and CoCos, the CoCo module adds a ‘coco’ object to the context. It’s through this context that the data passes between the CoCo itself and the rest of the conversational flow. The context object is part of CoCo's “context-transfer” standard for transferring data between bots, components, and live agents. In it you can find pairs of key-values relevant to the conversation and the component’s operation. Some of these keys are guaranteed to be present in the “updated_context” field when the component reports componet_done. You can find out which keys you get from each component on the CoCo Marketplace.
A best practice is to use the CoCo module as part of a subflow and just add handling for termination signals. Then, it’s simple to give a ‘switch flow’ command, which the subflow will route back to the main flow when the component has finished.
For more detailed instructions on how to use Cognigy.AI and CoCo, please see CoCo's guide on GitHub.
CoCo's vision has been to create a global index of conversational components where developers and conversation designers can share their work regardless of the platform they use. In order to do this, they have identified the top bot-building platforms in the world so that they could offer developers from each of these platforms a way to contribute components and/ or implement components that others have created on different platforms (such as DF, Rasa, Watson).
AI Conversation Systems