Home / Blog / AI is becoming a marketing word and it’s a pity -VI

AI is becoming a marketing word and it’s a pity -VI

A Chatbot is a typical example of Artificial Intelligence. It uses both Machine Learning and Deep Learning (a technique to achieve Machine Learning and extract meaning). Equipped with this technology a chatbot can engage in interaction using natural language, perform predictive analyses and sentiment analyses that in turns can lead to finely tune the interaction with the human. Image credit: Smartsheet

Chatbots

Chatbots, robots that can talk and interact with people using natural language are a reality and indeed they are a concrete example of artificial intelligence application. As shown in the figure they are able to learn, get trained to be able to speak and interact at two levels: to understand the person talking to them and to understand the topic the person is interested in.

The numbers around call centres are staggering, there are an estimated 3.4 million people working at call centres just in the US with a worldwide revenue exceeding 310 B$. That’s serious money!

Chatbots are supposed to flank these people (at least at first) to improve the perceived quality and to decrease the cost (also stated s improving the operational efficiency meaning no longer flanking but replacing people). According to Gartner, chatbots have been used in some 2% of customer interactions in 2017, expected to grow to 25% at the end of 2019.

As shown in the figure, they are clearly AI application. Hence, no hype there? Well, let’s say that in many cases, actually most cases today, the chatbot is a first level interface showing very limited intelligence, both in understanding the human and in delivering the answer. It is a bit smarter than an IVR – Interactive Voice Response- telling you press 1 if you…. (or say “1” if you…) but still a long way to intelligent. Most chatbots operate on a predefined decision-tree logic “if this then that” and their understanding of the human is, in most cases, tied to recognition of specific keywords.

There is a lot of potential and the huge market is clearly pushing towards better and better natural language support through the use of artificial intelligence. As mentioned, this is evolving through better understanding of natural language and through better knowledge of the problem space to be able to respond appropriately.

Chatbots are being used in many areas beyond call centres, including finance, healthcare, Insurance, real estate, hospitality. Image credit: Drift’s 2018 State of Chatbots Report

Interestingly, the evolution towards better understanding of natural language is also requiring sentiment analyses, i.e. to be able to detect the person’s mood. I was just talking yesterday to a person that used to work in a call centre of a major telecom operator in Italy and she told me that she quit the job because most of the calls required facing some very angry customer and interaction with them was really difficult. I mention this to stress how difficult it is for a chatbot to replace a human keeping the same level of customer satisfaction since this involves feeling what the human feels and resonate with that feeling through the interaction.

Notice that chatbots are applied in a growing number of fields, well beyond call centres. As an example, industry is considering them on the manufacturing floor, as a more effective way to interact with machines (robots) and processes.

For a list of some, at least to me, unusual application take a look at the Wordstream blog in its article presenting 10 of the most innovative chatbots.

Among them:

  • Endurance, a companion for people affected by Alzheimer
  • Casper, helping insomniac getting through the night
  • U-Report, created by UNICEF to give voice to marginalised communities
  • MedWhat, accelerating medical diagnoses.

In spite of their growing presence there is still quite a way to go before they can really be taken as companion au pair with us supporting an engaging conversation, even in a well defined area. Progress in sentiment analyses will be crucial to make us feeling we are talking with a person like artefact.

About Roberto Saracco

Roberto Saracco fell in love with technology and its implications long time ago. His background is in math and computer science. Until April 2017 he led the EIT Digital Italian Node and then was head of the Industrial Doctoral School of EIT Digital up to September 2018. Previously, up to December 2011 he was the Director of the Telecom Italia Future Centre in Venice, looking at the interplay of technology evolution, economics and society. At the turn of the century he led a World Bank-Infodev project to stimulate entrepreneurship in Latin America. He is a senior member of IEEE where he leads the Industry Advisory Board within the Future Directions Committee and co-chairs the Digital Reality Initiative. He teaches a Master course on Technology Forecasting and Market impact at the University of Trento. He has published over 100 papers in journals and magazines and 14 books.