“While customers can choose different modes of contact—touch (in a branch), type (on the website), tap (chat banking on WhatsApp) and talk (call the contact centre and interact with the voice assistant)—we see an increasing preference for digital self care,” said Anjani Rathor, chief digital experience officer, HDFC Bank, pointing at the increase in such interactions. Today, a significant part of the 30 million-plus monthly customer interactions at India’s largest private sector bank are processed by AI/machine learning (ML) solutions, including Eva.
At white goods major Voltas, half of the customer engagements, including demo, installation and service requests, are now handled by bots. This is a massive shift from 2018-19, when 95% of customer calls were handled by human agents. At Japanese home appliances and consumer electronics major Panasonic India, bots handled 20% of the queries when they were introduced in 2020. Today, that has doubled.
The world of the contact centre, once manned by an army of humans, has been transformed by technology. With every new technological advance, from interactive voice responses (IVR) almost three decades back to chatbots a decade or so ago, the human army has become a little smaller, and the promise of better engagement between customers and companies has become bigger.
The outsourced contact centre industry, pegged to be worth $110-115 billion globally at the end of 2023, will contract to $95-105 billion in the next three to four years, according to Everest Group.
At first, it was patchy, good enough only for ‘routine’ tasks such as resetting a password, ordering a cheque book or updating an account’s balance. That left all the stakeholders—the company, customers, and the human agents—wanting more. Complaining customers fret about talking to chatbots who can’t comprehend their ‘intent’. Human agents, on the other hand, have to turn the other cheek and endure the ire of customers (and frequently, their abuse) to ensure the company they represent appears to put the customer first.
But today, customer engagement is dramatically changing thanks to AI. Generative AI (GenAI) powered chatbots are far better in comprehending and processing human languages and addressing customer needs than their earlier avatars.
So much so, that in a 24 April interview with The Financial Times, K. Krithivasan, chief executive officer (CEO) of the $30 billion tech services major TCS, said that AI will result in a “minimal” need for call centres within a year. “Chatbots will soon be able to analyse customers’ transaction history and do much of the work done by call centre agents,” he noted.
View Full Image
But there is a dark lining on that silver cloud. The outsourced contact centre industry, pegged to be worth $120-130 billion globally at the end of 2023, will contract to $95-105 billion in the next three to four years, according to Everest Group, a research and advisory group specializing in the global services industry.
It’s not as if the volume of work will decline; but bots, which cost far less than people, will take on more of the workload. Simultaneously, human agents will become more efficient, and focus on more complex tasks, such as helping customers understand insurance policies, reschedule flight bookings, and so on. And enterprises, consequently, will be able to reduce their overall spends on contact centres.
“Five years back, calls were the primary means to support customers—handling 70% of the queries. Now, WhatsApp, chatbots, and voice bots work alongside human agents. At least 30% of the contact centres globally have adopted AI-powered tools,” said Ishwar Sridharan, co-founder and chief operating officer (COO) of Exotel, a company that offers AI-driven contact centre solutions.
Bullish on bots
View Full Image
The conduit between the company and the customer has evolved over the decades from face-to-face interactions to telephone calls. When the call volume ballooned it led to the evolution of call centres manned by a sea of headset-wearing human agents. With the internet, these centres could be set up anywhere. Airtel Ola, Uber, Tata Motors, Bank of America, Swiggy, Dell, JP Morgan Chase, HSBC, Capital One, American Express, P&G and hundreds of other companies either run in-house divisions or outsource their contact centre work. Globally, more than 12 million are employed in the industry today, answering customer queries for international as well as local customers. A million or so of them are in India, which has become the world’s call centre thanks to its large population and low costs.
The Philippines, with around 1.2 million workers, is another large global contact centre destination. But those numbers could decline with the emergence of the chatbot, which is evolving rapidly since being initially deployed to handle basic customer queries. “Contact centres will grow more on the digital side than on the human labour side. Average call handling times will reduce as bots get better in comprehending caller needs,” said Tom Coshow, senior director analyst, Gartner.
“The (call centre) industry started with answering basic queries like credit card outstanding, flight schedule, password reset, etc., which is now mostly automated,” noted Raman Roy, chairman and MD of Quatrro BPO, a pioneer in India’s back-office industry. “In six-eight years, technology will take over much of the high-end work. Better natural language processing is making it possible.”
Contact centres will grow more on the digital side than on the human labour side
—Tom Coshow
It has already begun to do that and is helping companies save huge sums of money. Swedish fintech company Klarna, is a case in point. In 2023, Klarna partnered with OpenAI to develop a virtual assistant. This March, the fintech claimed its virtual agent helped shrink its query resolution time from 11 minutes to just two. The assistant does the work of 700 humans and Klarna expects to save $40 million this year.
How the bots evolved
About a decade back the early chatbots searched for keywords in text (like, reset password or bank balance). While they could do frequently asked questions (FAQ) kind of work, the experience was not too great for anything that required a transaction. But it did free up human agent time to focus on more complex tasks.
Then came intent-based bots—if a customer asked for a red coloured Air Jordan shoe, the bot knew the customer was looking for a shoe. The bots got a boost as the world shut down during covid and companies scrambled for digital alternatives to answer customer queries.
View Full Image
Now, a GenAI version chatbot can understand out-of-context text. So, if a user asks, “I have knee pain, which shoe should I wear,” the bot will be able to offer options.
“Contact centres have problems with hiring, training, attrition and customer satisfaction. With chatbots these problems are getting resolved,” said Ganesh Gopalan, co-founder and CEO of Gnani.ai. Set up by a group of former Texas Instruments engineers, Gnani.ai is a conversational AI platform backed by Samsung Ventures.
While chatbots are improving with better natural language processing capabilities, lower costs also work in their favour. According to industry estimates, the average per-employee cost of a contact centre ranges between ₹30,000 to ₹45,000 per month. Each agent does around 1,000 calls a month at ₹15 to ₹20 per call. Chatbots, on the other hand, are able to deliver at ₹3 to ₹5 per call.
Bots today are learning both from historical conversations fed to them and new ones they attend to in real time. “The underlying technology is AI now. There’s more intelligence built in chatbots and they can even give a recommendation based on a combination of knowledge and intelligence,” said Rashid Khan, chief product officer and co-founder of Yellow.ai. His company focuses on automating access—be it voice, chat or email. Backed by Salesforce, WestBridge Capital and others, Yellow.ai claims to have over 1,100 customers, of which 400, including the Bajaj Group, Asian Paints and Tata Power, are in India.
Interestingly, there are more ventures like Exotel, Gnani.ai and Yellow.ai being set up today by entrepreneurs than call centres, as technology is tilting the scales in favour of chatbots. The call centre wave ended around the time the chatbots started appearing. It made way for an omnichannel customer experience that didn’t rely solely on humans but also included chatbots and social media platforms such as WhatsApp.
Impact on jobs
View Full Image
Bots are getting better but humans won’t entirely disappear from contact centres, at least in the foreseeable future, for reasons spanning from regulation to technology limitations. “While technology is improving, regulation could delay a blanket use of bots,” said Quatrro’s Roy. Contact centres provide employment at scale and governments could step in to check a complete wipeout, at least, in the short term.
Besides, European and American regulators are coming up with guidelines on how companies portray chatbots to website visitors, users and customers. California’s laws restrain companies from making their chatbots appear as real humans. European regulators bar the approval of consumer loans by chatbots.
Some sectors like banking may not adopt GenAI in a hurry.
“We are custodians of customer money and data. A cautious approach on GenAI, ensuring data privacy and security, is the right thing to do. A good part of our innovation technology is in-house,” said HDFC Bank’s Rathor. Also, banks see themselves evolving with deeper customer engagements on, say, wealth management and insurance, while bots take care of routine queries.
California’s laws restrain companies from making chatbots seem like real humans. European rules bar the approval of consumer loans by chatbots.
Even as regulators step in, chatbots can’t take over the entirety of contact centre operations. “You can’t reschedule a flight via a chatbot,” explained Gartner’s Coshow.
“Data privacy, security, and hallucinations (GenAI giving out incorrect or nonsensical information) are among the major challenges with GenAI-powered bots. We will still require the human touch in this industry,” said Sharang Sharma, vice president, Everest Group.
Roy pointed out that processes such as mortgage origination have 1,800 questions. “It’s hard to say that it will all get automated,” he added.
On their part, contact centre companies believe their future lies in a blended model. “We use technology not to replace humans but to expand their capabilities and accelerate resolution through conversational chatbots for simple queries. We use our team of interaction experts for more complex resolution,” said Teleperformance India’s COO M.V. Prasanth.
Paris-headquartered Teleperformance has 500,000 employees globally and around 90,000 in India, serving more than 200 clients across industries.
Swedish outsourcing company Transcom acquired Gurgaon-based contact centre VCosmos in February to expand its operations here. “Natural language processing augments our capability to offer services to any market. Improvements in the technology will be key to automating processes, but humans will be needed for personalization,” said Amandeep Singh Arora, chief experience officer and managing director, Transcom India. Transcom Global has 30,000 employees in 30 markets, of which 700 are in India.
Key to adoption
Even as the call centre business stares at its most disruptive moment with the proliferation of chatbots, customer satisfaction will be the key to quick adoption. “Using only humans is suddenly seen as an inefficient way to run a contact centre,” said Gaurav Vasu, CEO of Unearthinsight, a Bengaluru-based consultancy. “Omnichannel models, with human capability augmented by intelligent bots, will get strengthened.”
Those running contact centres, meanwhile, will see margin improvements with bots, compelling them to use more of them. Businesses will tilt towards technology as these (customer care divisions) are seen as ‘cost centres’, which won’t add to their bottomline.
According to Unearthinsight, pure human voice business gives a company a 9-11% margin; humans combined with bots and email support is 12-13%. Adding maintenance will widen that to a 14-15% gross margin. But while companies get bullish on bots, customers are likely to react in different ways. Some, especially those from the older generation, will always prefer to interact with a human—and feel a deep sense of frustration interacting with a bot. “When my mother calls up Microsoft tech support, she expects Bill Gates to answer! For my teenage daughter, an exchange with a chatbot is good enough,” said Gartner’s Coshow.
When that teenager gets to her grandma’s age, she may not even remember that an army of humans once populated contact centres. To be sure, they’ll still be around, just not so many of them. Indeed, contact centres will increasingly become chatbot centres. But that also means there’s likely to be a less-harassed human being waiting to answer that call when it is beyond the chatbot’s capability. And that’s a win-win for all.