Artificial Intelligence

Chatbots and Virtual Assistants. Will you Know who’s Talking?

13/12/18 3 min. read

Do we really know who’s speaking when we chat with a customer support centre? Are you absolutely sure that it’s a real person? Gartner predicts that by 2020, virtual assistants will handle 85% of customer service and support operations.

This is not only a surprising figure but also an imminent reality. We’re certain that all of you have heard about chatbots and virtual assistants and most likely you’ve interacted with them already. What is a chatbot? And why does it seem like everyone is talking about them lately?

What is a chatbot?

A chatbot is an IT system capable of holding a conversation with a human being using natural language. Chatbots are particularly salient insofar as their immediacy of response, adherence to the brand image and capabilities to respond several clients simultaneously. Chatbots, also referred to as conversational assistants, provide companies with yet another interface to interact with real and potential clients and even their own employees

What are their benefits?

  • Simple interfaces and very short adoption times: since we are already accustomed to chatting across different messaging apps, these tools are very easy to use.
  • 24×7 availability:chatbots are always available and can even redirect conversations to other channels attended by real people.
  • Orderly responses: chatbots are trained to give responses that are in line with company policies. Moreover, they don’t respond what they don’t know.
  • Responses across multiple channels: ranging across websites or mobile devices, including Contact Centre IVR (Interactive Voice Response) and even channels like Telegram, Whatsapp or Skype.

Where are Chatbots used?

They can be used virtually in all sectors and services: customer support, sales, private areas, intranets, etc., and even in specific areas such as Human Resources to interview job candidates.

Chatbots can be useful in banking to let users know their balances, execute transactions, send minimum balance alerts or send commercial product recommendations.

They nevertheless require a bit of care to avoid of what happened to Tay, Microsoft’s chatbot, who began to spew out racial slurs after learning from the users who were interacting with it.

What kind of technology do they use?

Chatbots face two main challenges: understanding the question and then formulating the answer.

Once they detect what the user wants to know, i.e., their intention, the answer must be formulated by parsing the questions for possible meanings or creating dialogues to guide the user to the correct answer for their initial doubts or concerns. This seems like a very simple challenge, but it’s actually a very involved process that requires a great deal of effort.

And what are the legal issues?

One shouldn’t overlook the vast legal dimensions associated with chatbots, including aspects such as information received from users (personal and/or financial data), industrial property or contractual responsibilities and liabilities when answering them.

Get deeper on knowing Bots with examples: How to teach a Bot to talk: examples

Author: Ernesto Rubio & Marta Sanchez

Santander Global Tech



Other posts