Originally launching as a chatbot analytics platform in 2017, Area 120’s Chatbase can now design virtual agents that leverage machine learning to be more helpful, and over time improve. The new Chatbase Virtual Agent Modeling service is a part of Google’s internal incubator and is available in early access.
In the early days of voice and virtual assistants, chatbots were heavily touted as the next big thing, with customer service now seen as the biggest use case. At I/O 2017, Area 120 introduced Chatbase to analyze and optimize virtual agents used by third-parties to help consumers.
Built within Google’s internal incubator, Chatbase is now leveraging what it has learned after 18 months of analyzing hundreds of thousands of bots and billions of messages.
Chatbase Virtual Agent Modeling can look at a company’s customer transcripts — at least 100,000 English-language live-chats are required — to create categories, and find specific help requests. Chatbase then models a detailed conversation flow that reflects how customer interactions play out.
For complex intents, Chatbase models simple yet rich flows developers can use to build a voice or chat virtual agent that handles up to 99% of interactions, responds helpfully to follow-up questions, and knows exactly when to do a hand-off to a live agent.
In practice, Chatbase argues that its analysis can cut “weeks, months, or even years from development time.” Results can be exported to virtual agents like Dialogflow.
Chatbase’s Virtual Agent Modeling is available as an Early Access Program for interested customers. It’s an expansion of how Google Cloud is offering machine learning services that big companies can leverage and integrate.