Write to us

Contributors

Writers

Priscilla Thomas

+

A content writer who has worked for Infosys Technologies and other technology startups in India and abroad.

Moulishree Srivastava

+

A freelance journalist with over 7 years of experience, she writes research-based analytical stories on technology and business.

Avanish Tiwary

+

Avanish is a Bangalore-based journalist who writes on business with a specific focus on technology companies.

Priyanka Bhatacharya

+

Priyanka has covered every aspect of the IT industry as a tech journalist since its early days. She is now an independent writer, working on subjects like digital marketing, enterprise technology and high-performance computing, among others. 

S.Sahu

+

A freelance content writer, S. Sahu was the former editor of TCS's house magazine at Tata Consultancy Services. He developed tech marketing collateral for the company and helped compile and edit books and journal articles on TCS's technology innovations. He also ghostwrites print and online publications. 

Prajwala Hegde

+

Prajwala is a Bangalore-based freelance journalist who writes on social issues, stories of human interest, and art and culture, among others. 

Rajesh Nanarpuzha

+

Rajesh Nanarpuzha is an Assistant Professor of Marketing at IIM Udaipur. Previously, he has worked as a brand manager in Dabur, and as a business consultant in the retail and consumer goods domains at Cognizant and Tata Consultancy Services. Rajesh has an MBA from IIM Indore and a doctorate in marketing from IIM Ahmedabad.

Designer

Priyokumar Singh Naorem

+

He is a passionate UI & UX designer who thrives on creating engaging creative solutions.

Artists

Dyuti Mittal

+

A freelance illustrator, artist, graphic novelist and designer. She has designed and illustrated several book covers. Her personal illustrations so far have attempted to seize the fleeting absurdity and mood of places, things and people she encounters in a childlike, intuitive and expressive manner with closure, beauty and innocence – the things that she desires.

Sumakshi Singh

+

She is an artist and an educator who has taught and lectured at The School of the Art Institute of Chicago, Oxford University, the Victoria and Albert Museum among other institutions. Her installations, paintings, thread work and sculptures have been exhibited in Saatchi Gallery - London, C24 Gallery - New York, and Museum of Contemporary Art – Lyon, among other notable galleries and museums from around the world.

Parul Gupta

+

A commerce graduate from Delhi University, Parul pursued a masters in fine arts from Nottingham Trent University in the UK. As an artist, she is interested in line as a subject which has led her to follow architectural lines in built environments. She says she is also interested in how we perceive the environment that we inhabit and what happens when a subtle shift is made in things which we have been used to seeing in a certain way. We present six of her artworks here.

Shweta Malhotra

+

Shweta Malhotra is a graphic artist and designer from Mumbai, based in New Delhi.
After working with ad agencies and design studios for close to 8 years, she branched out on her own and currently works independently.s Her overall design aesthetic is minimal, bold and graphic, a response to the maximalist visual language prevalent in India.

Rithika Merchant

+

Rithika Merchant (b.1986) received her Bachelor’s Degree in Fine Arts from Parsons the New School for Design, New York in 2008. She has exhibited extensively since her graduation. Recent exhibitions include a duo show “Reliquaries: The Remembered Self” at TARQ, Mumbai; “Language of the Birds: Occult and Art” at 80WSE Gallery, New York; and group shows at Summerhall, Edinburgh and Artry Gallery, Kochi. Her work has been included in multiple group shows at Stephen Romano Gallery and The Morbid Anatomy Museum, New York. Born in Mumbai, she now divides her time between Mumbai and Barcelona.

Aniruddh Mehta

+

Aniruddh Mehta is an artist based out of Mumbai, India. Trained in graphic design from the London College of Communications, Aniruddh is a self-taught illustrator and currently works as an independent freelance designer. He believes in finding the right balance between art and graphic design. He has worked closely in collaboration with Bhavishyavani Future Soundz, Qilla Records, Taxi Fabric, Adidas, Dell and United Colours of Benetton. He also goes by the moniker, ‘thebigfatminimalist’ and his style ranges from bold minimal forms to more intricate pieces exploring patterns and geometry.

Paramesh Jolad

+

Paramesh is an artist who enjoys working in both the realistic and abstract style of painting. He loves working with water color. Featured in this issue are a set of water color works that he has created exclusively for us on the subject of digital transformation.

Chandrashekhar Thakur

+

Based in Mumbai, Chandrashekhar Thakur is a Senior Art Director and Illustrator at Truebil.com and the Founder of HAPPiNESS For You. He loves working with new styles of art and considers illustration to be his forte. Chandrashekhar has completed his BFA from DY Patil College of Applied Art.

Concept and Direction

Nimish Vohra

+

Head of marketing at Regalix, Nimish drives research in emerging technologies and customer experience, and takes a keen interest in creative arts.

Enquiry

+91 9560509289

aishani.majumdar@regalix-inc.com

Know when a new article is published

As messaging emerges as the preferred mode of communication in today's world, many companies are leveraging the trend to serve their customers better and faster, without having to set up a call center. Online restaurant search and discovery platform Zomato, is one amongst them.

Since its inception, Zomato has been using chat as a primary and only means to offer customer support services. A couple of months back, it moved its chat-based customer service in-house. The company has started investing in a chat platform as it plans to evolve its customer support system. Over the next few years, Zomato sees artificial intelligence (AI) driving chat and natural language processing (NLP) becoming a common tool to assess the quality of customer service provided by its agents, says Pradyot Ghate, Associate Vice President and Product Lead at Zomato, in an interview. 

 

Interviewed by Moulishree Shrivastava

 


 

/ / How has the nature of customer support changed or evolved over the last few years?

 

Primarily what we have seen evolve over the last couple of years is the mode of communication that people have adopted. If you look at 5 to 10 years ago, the phone was the primary mode of communication. People used to call each other if they wanted to talk. Now over the years, people have moved to chat a lot. They WhatsApp each other, instead of a call. 

Calling requires both the parties to be available at the same time. Whereas if we are connected through WhatsApp, you can shoot me a question, and I can answer you when I am available. The process can be asynchronous. As a result, the adoption of messaging technology has been far higher across all the age groups. 

When we started online ordering, the problem we were solving was that people didn’t want to talk to restaurants over the phone to place an order. We said, if that is the problem we have successfully solved, then why should we expect them to call our customer service for help? These people are averse to phone calls, and they prefer using an app to communicate their orders. That is why from day one we have adopted chat as our primary mode of communication in the B2C space. And it has been very successful. We have scaled to millions of orders a month today. We have never had the need or demand from the user side to open up a helpline number. They are comfortable chatting with us.

 

/ / What are some of the technologies you have adopted to transform customer service? 

 

We knew we need to get into Natural Language Processing (NLP) and Machine Learning (ML) to facilitate better communication with our users. The first step that we have taken is to move the chat system completely in-house. So the core technology that powers the in-house chat engine has been custom-built to cater to future needs of ML and NLP assessments. Earlier we used to use a third party engine. Now we have built our own from scratch. 

If you look at it, there are no successful chatbots today. If you speak to some of the chatbots which are available on Facebook Messenger, they don't feel AI-powered - you do not feel like you’re talking to a human being. Most of the chatbots feel like you are filling in an HTML form and they are replying to you based on the answers you have entered. So I think, on the chatbot end, the technology has not really matured yet. It will take some time to get there. 

We have started with some early experiments on NLP in Zomato. For example, we have tried to look at our chats in real time and assess the mood and temper of the person who is talking to us. These are all beta technologies as of now. We are trying to assess if we can understand the sentiment of the caller. The other thing we are trying to do is use an NLP algorithm to quality check the answers that are being given. One of the things we are very particular about is to make sure that the quality of our customer service remains high. To ensure that, until today, we have been doing human checks. We sample a couple of chats, and our quality team goes through them, and gives their feedback on what could have been done better. So what we are looking at in the next couple of months is to automate this process, so quality check can be done not just on a sample of chats, but on all chats using ML algorithm.  The algorithm can then give a feedback to the agents on where the chats went well and where there is scope for improvement. We are trying to build a system wherein at the end of the day when the agent logs out, he gets a feedback report, so the next day they can do a better job at serving our customers.

 

/ / Do you think few years down the line, chat-based customer support system will be completely automated, and human agents will only be pulled in as and when the need arises?

 

It will definitely happen down the line, but it’s fuzzy as to when it will happen. We are starting to understand how we construct and deconstruct sentences, and how we maintain context between one sentence and another. We have complex ways of understanding and using language. And we have multiple ways of saying the same thing. We could say something with sarcasm, for example. There, the context of the conversation has to be kept in mind. If I am not happy and I end up my conversation by saying, I like Zomato - that would be sarcasm. But an NLP algorithm at present won't be able to understand that. The system might reply with ‘thank you’, which is a very bad move. We, by which I mean the whole product tech community, are still far away from passing the Turing test - when you wouldn't be able to differentiate between whether you are chatting with a human or a bot. 

 

/ / What are some of the trends that you see emerging in customer service?

 

Three years ago, when Facebook announced that they were going to open up messenger for chatbots, there was a lot of excitement, but over the years, that excitement has sort of died down. Multiple companies came up, got funding to make chatbots, and they never moved past series A or series B round. And today there are really far and few chatbot companies left that are sort of working through things and trying to make chatbots. But I think in future, people are going to move chatbots in-house because each chatbot will have to be custom-built for each use case. For example, if you are an insurance company, your chatbot has to be specialized in talking about insurance, and how to sell it and how to help people claim insurance and resolve queries. Similarly, in case of Zomato, we will specialize our chatbot to handle queries in the restaurant and food business. Just like humans have specializations, I believe chatbots will also have specializations and they will be custom-built for different situations, considering organizational policies and how things work internally.

Over the next few years what would happen is that the agents will get exposed to the chatbots first. Let’s say a customer texts something to an agent. The agent will see if there are options (as suggested by the chatbot) that he can reply with. He will either pick one of them or write something else on his own. And this will be the learning that the chatbot will go through. This will be a sort of child-parent relationship, wherein the child suggests something, and if the parent does something else, the child will learn from that behavior. That's the way chatbots will evolve. As we gather more data, we will be able to completely automate it.

 

/ / What are the challenges that you face as you try to adopt technology to serve your customers better?

 

Thankfully our customers are tech-savvy. Whenever we come up with a new technology, they are very quick to adopt it and provide feedback. The critical function of adopting any new tech or offering new tech to customers is to constantly get a feedback on how they feel about it. I think what we need to be very vigilant about is how exactly the new tech is evolving, if it is useful for customers, and whether or not we have to retool it to serve them better.

 

/ / In providing chat-based customer services, is language also a challenge? 

 

Today we are present in 23 countries. We support 13 different languages across the world. And the Zomato app is available in 13 languages. But the chat is centralized in India and is supported by our Indian team. At present, our chat function only supports Roman characters. We are cognizant of this fact; yes, we have to become more vernacular in India and more international language oriented for the rest of the world. We have been working on this.