Educating robots the art of a good conversation
BENGALURU: it pulled the plug on it within a week When lingerie startup Clovia deployed a bot to talk with clients online two years back. A customer asking about an merchandise was shown bras to purchase. Gaffes like this can spiral out of control.
And yet after some persuasion, Clovia decided to try out a chatbot. From throwing a bot this time the strategy was a lot different.
The first few weeks have been spent digesting all available data on client queries, understanding business processes, identifying areas that were easiest to automate first, and training the bot to deal with those and hand off trickier areas like body contours and matches to humans. Subsequently Clovia felt confident enough to provide another try to the chatbot. There has been a reduction in people handling customer queries even though the business of Clovia has increased. Interactions with the bot are increasing in number and sophistication.
“Originally we gave it very little space, but as we saw customers responding better, we have begun to push the chatbot further,” states Vermani. “Many customers are more comfortable talking to a bot about something as intimate as a lingerie. We want an bra salesperson to notify a consumer. And we are pretty much there in our test runs.”
NOT JUST A CHATBOT
Many people have interacted with chatbots only to be frustrated whenever the conversation moved beyond basic questions. That’s because most chatbots do not have any idea about the intent behind our queries. The possibilities of natural language processing (NLP) have exploded over the past decade as the big tech companies-Google, Amazon, Microsoft, Apple, Facebook and IBM-have pumped resources into creating their own virtual assistants such as Alexa and Siri in addition to open source AI platforms. This has lowered the entry barrier for producing chatbots but has not made them intelligent in specific contexts. A baby doll means something in the toy market and something in the lingerie of women , for example. “Everybody says,’Oh, I can make bots’. That is like saying I play cricket and so does Virat Kohli. Building a bot that is simple is college level material. As soon as you enter a higher level, data sciences and business use cases come into play.”
Satish Medapati, who managed the APAC region 7. Ai turning entrepreneur, started Intentico in February 2018. The key, according to him, is to look at the chatbot as one piece of an enterprise’s”AI transformation” instead of a plug-and-play tool.
“Having the best algorithm or NLP is not enough. A whole lot of consultation work also needs to be done,” says Medapati.
Only then can a chatbot be incorporated into a roadmap to automate and enhance the experience of today’s customers who are used to messaging and anticipate relevant answers immediately rather than being put on hold in a call with repetitive music.
Intentico is focused on drilling heavy with a few enterprises large enough to architect such a job rather than spreading itself thin across multiple smaller customers. Automotive, banking and retail firms in India and Japanese electronics manufacturer Toshiba in the Gulf are one of its early adopters. “Each of our clients has failed earlier with bots despite a multi-million-dollar exposure,” states Medapati.
Intentico’s NLP motor Curie in the rear end and its text and address bots at the front end are part of a larger suite of products that deal with various facets of customer interaction across multiple channels. These include bot-integrated middleware to connect different systems and stations,’botlytics’ to derive actionable insights, and traditional tools like IVR (interactive voice response), ticketing and dashboards.
Such an integrated strategy increases the odds of its chatbots engaging in deeper discussions and making sense. Intentico works with partners and system integrators to locate and fit all the pieces of the puzzle. In the case of Toshiba in Dubai, this extended to figuring out the infrastructure for messaging since many countries in the Gulf have banned VoIP (Voice over Internet Protocol) which makes popular programs like WhatsApp inaccessible. Intentico crossed that hurdle to station an whole call centre’s customer queries to a language bot. “More than 95 percent of the queries are resolved with the Tosh speech bot,” says Medapati.
How Intentico has built its NLP and flow engine Curie is also different from the decision tree model that most chatbots use. “Ours is an ensemble of decision trees in which the bot can leap from one tree to another,” says Medapati. This is closer to the way the person conversation flows. By way of example, a Toshiba customer might start a conversation about a service request and suddenly ask about a product she found at a retail store. “The human mind can switch seamlessly from one context to another but it is very difficult for a machine to perform,” explains Medapati. “What we have allowed in Curie is a context switch mechanism in which the bot can jump from the node of one tree to a node in another tree of the ensemble.”
Intentico is the third startup of Medapati. His earlier venture Bounty needed a proprietary indoor positioning system to monitor a client’s location in the outlets of partners. It engaged customers by demonstrating reward points to collect and generated client analytics for partners.