Conversable launches bots that learn better customer care from employees


Austin-based artificial intelligence company Conversable Inc. today announced the launch of its bot platform AQUA that’s intended to help employees improve customer service more easily.

The bot, which stands for “Answer Questions Using Artificial intelligence,” allows businesses to deploy chat bots that can answer customer questions. And when questions cannot be answered, employees can update the bots in real time without the need for programming experience.

At its core, AQUA is a machine learning platform that acts as a bridge to customer service by fielding questions from clients. When the system knows the answers to those questions, the usual thing happens: The bot answers and the customers goes on their merry way. However, when a question exceeds the bot’s capabilities, the bot hands it off to a human employee, usually a customer service representative. When this happens, the AQUA system learns from the chat between customer and employee, so that this happens less and less often.

Where the system really begins to shine is that commonly problematic or complex questions are noticed by AQUA and pushed up the pipeline in a dashboard for higher-level employees such as project managers and experts. At this point, an employee looking at the dashboard can interact with these queries and directly modify the answers the bot would give or the information relationships the bot uses to deliver answers. The answer the employee writes is then integrated back into AQUA’s conversation trees without the need for re-programming or AI expertise.

“The AQUA system combines human-assisted machine learning with adaptive systems and an easy to use console to let business managers deliver answers quickly, easily, and correctly, to their customers,” said Andrew Busey, co-founder and chief product officer of Conversable.

Traditional AI bot setups require time from Information Technology teams in order to update conversation flowcharts to add answers and increase the knowledge base of the bot. The AQUA system does away with as much of that need as possible by making it easy for employees to get on with business.

Another way that AQUA makes it easy for business managers and non-technical employees to bridge that gap is the ability to chat and modify AQUA from the dashboard as well, Conversable Chief Executive Officer Ben Lamm told SiliconANGLE.

Image: Conversable

Image: Conversable

Project managers and other non-technical experts can open up a sort of “debug” chat with the AQUA system and ask questions as if they’re a customer with a problem and watch what happens. If the bot’s answers go askew, the expert can then add to the knowledge base right at that point, giving the system a better idea of how to answer the same sort of question should it arise in the future.

With any business there are a huge number of questions that can arise when it comes to customer care. Automated systems, such as customer service bots, can field a large set of common, simple questions but complex, less common questions make up a significant number of questions that take up a lot of customer service time. With AQUA, each time one of these uncommon questions is encountered and answered by a human representative, the system gets smarter reducing that time. Also, uncommon questions pushed up the pipeline as insights to business managers further reduce time spent on customer service, meaning quicker answers for customers.

Conversable’s platform powers a lot of customer care across various industries, including a recipe finder for Whole Foods, ordering and reservations for TGI Fridays, order support for Pizza Hut and Wingstop, and a live event bot for Viacom and MTVThe platform provides bots that can run across numerous platforms including Facebook Messenger, Twitter, Short Message Service mobile texting, and Amazon Alexa.

Image: Conversable