[PRESS] ChatBots theory and case studies

During the Workshop “ChatBots theory and case studies”, last Tuesday in Turin, we discussed about how machines and new technologies will be able to artificially reproduce intelligence and adopt “rational” behaviors, give answers and take decisions based on the interaction with human beings, on the interpretation of the context and on the recognition of physical elements.

What is a ChatBot?

A ChatBot is a software based on AI, able to simulate an intelligent conversation with users on a chat system. It offers a functional support service through the main messaging platforms such as Slack, Telegram, Facebook Messenger.

What are the possible uses?

ChatBots can be used for different purposes like:

  • Customer care
  • Product Suggestions
  • Offers and promotions
  • News & Weather
  • Education
  • Fun

The main strength of ChatBots is that they are always available to offer users help and answers.

And at the same time they can trace users’ interests, preferences, ages and tastes.

Michele Papaleo, from IndigoAI, showed us how to use ChatBots and AI to improve Customer Experience, since managing effectively the Customer Service is a decisive element for success.

The Customer Experience thus becomes a milestone on which the success of the company depends on.

It is fundamental to improve (methodologically, but also through the appropriate technological choices) the overall experience that customers will “live” throughout their relationship with the company itself: Artificial IntelligenceChatbot and Virtual Reality will decrease the weight of human interaction in favor of a more immersive digital Customer Experience, a phenomenon that companies must take into account in redefining their Customer Service.

The language process, that is at the basis of ChatBots, belongs to various coexisting and cooperating schemes, currently still largely to be discovered. In fact, they are not able to follow complex semantics at this time. For example, they can not venture into metaphorical dialogues, highlighting the first limit of their functionality.

However, we know that once a scheme is developed, a robot can execute it.

But how?

The Xenialab team showed us two demos.

– Hey Boss, how are you?

– How many times have you played Born to Run?

– How many times have you played in Turin? 

HeyBoss, passionate about Bruce Springsteen, answered our questions about the concerts of the great American singer-songwriter.

HeyBoss is developed using Google’s Dialogflow and choosing Facebook Messenger as an integrated channel.

The main strength of the ChatBots, as we can see from this demo, is to be a service always available to get instant answers to simple questions but…when complexity increases, Chatbots are not enough!

For this reason it is very useful to support, especially for customer care, agents and “real” operators to ChatBots.

If the ChatBot doesn’t have enough elements to handle the conversation, this is automatically forwarded to an agent that can continue it and provide the requested information to the customer.

One of these software solutions is the Omnichannel XCALLY suite, developed by Xenialab in Turin, used by Contact Centers in more than 50 countries: it is designed for customer care who need to communicate through different channels (Voice, Fax, SMS, Email, Chat, Social …).

The second demo has shown this concept through the Parrot ChatBot, developed with api.ai, integrated with the XCALLY software.

chatbots-xcally

The Parrot ChatBot is always ready to help tourists looking for information about Turin: hotels, museums and attractions…

If a tourist asks for a road indication? The Parrot informs the user to wait and puts him in touch with an available agent, who can manage the conversation via the XCALLY web interface.

The event ended with a series of brief but effective suggestions for those who approach the development of Chatbot: it is essential to choose which technology to adopt, to define a precise purpose, to follow a style in a consistent way, to design interaction always thinking about end user and … try, try, try!