LiveChat aiming at true AI with BotEngine launch
LiveChat, a chat application for customer service and online sales, with years of experience across different markets and industries, came best equipped to develop a modern mechanism based on A.I. Its goal is helping every business to find its own way to participate in the new reality of chatbots.
- In the past we seemed to be profoundly convinced that chatbots would be relevant only to tech enterprises, but today even small and medium businesses can start chatbot-powered conversations with their customers, even if they aren’t tech savvy - said Mariusz Ciepły, CEO at LiveChat - BotEngine’s chatbots can be configured for a number of situations happening during conversations between brands and users including ordering offered products, checking the availability of hotel rooms or canceling restaurant’s table reservation.
The chatbots are mostly used to boost the companies’ customer service. The return on investment from using the AI-powered customer service can come really fast and it’s mostly visible in cost reduction, efficiency and increased customer satisfaction.
Currently LiveChat clients use canned responses (a pre-made answers to popular questions that can be recalled with just a few keystrokes). In LiveChat’s case this function learns about 1000 phrases daily. The more agents use these pre-made replies, the smarter they are.
To date, 20% of all 1000 chats, managed daily by LiveChat’s Supports Heroes are solved with the canned responses suggestions. As a consequence the team’s efficiency has grown by over 50%, allowing agents to support not just 6, but up to 9 concurrent chats. LiveChats’ Support Heroes answers on average 25,000 questions monthly and, from the beginning of 2017, keep 98,5% customer satisfaction rate.
Having that in mind, LiveChat wanted to be one step ahead and offer companies a tool, which automates their customer service, making it even more effective and quicker. BotEngine’s mechanism is largely based on the canned responses learning process.
Even though BotEngine offers advanced features, its implementation in messaging apps is really simple and doesn’t require programming changes. The possible interactions between bot and user can be built directly in the application and shown as a tree structure. In this way it’s easy to imagine how the conversations would look like and improve it if needed.
During a conversation with a bot, ‘customer’s message’ is analyzed by BotEngine’s advanced NLP mechanism - so called ‘Queries matching’. The platform’s creators did their best leveraging the speed of bot’s answer and content analyzing time. The long-term goal for BotEngine development team is creating human-like bots. They strongly believe that they can only achieve this by machine learning process being powered by gathering conversation archives.
LiveChat was the first user of BotEngine and has prepared scenario of a debug process for its customers. The company plans to use it just after the integration of both platforms. Ultimately, the user will start chatting with a bot, but the conversation would be forwarded to one of the available agents if needed.
- Our goals for BotEngine are bold so we want to build our product with everyone who is an AI enthusiast. BotEngine already integrates with Facebook Messenger. Currently we’re working on integrations with LiveChat, Slack, Kik and other messaging apps, planned at the end of this year - says Mariusz Ciepły.
A strong community is being created around the platform with the aim to tailor the product to users’ needs. The community will enable developers to share knowledge, at the same time giving the opportunity to reach 20 000 LiveChat customers with immediate access to BotEngine.