How web chatbots are putting millions of low-level jobs at risk

Try asking her if she likes you. Her answer: “Am still learning.” Ask her if she loves you. “Stop! This is going too fast for me.” Try finding out what she’s doing tonight. “Oh, nothing much. I’m just putting the final touches on my plan to conquer the world.” Does she believe in God? “You have god, I have code.”

Eva’s cute and chirpy, and quick on the counter. She’s a web chatbot — a software robot designed by Senseforth, a botmaker, to simulate conversation with humans — employed by HDFC Bank to help visitors looking for products and services online. Barely a few months since going live, Eva has handled over one lakh customer chats. On busy weekdays, she is on “concurrent chat” with over 150 bank customers. “Eva is a ‘conversational agent’ and she helps us connect with our large customer base. For us, this is a technological leap to old conversational banking,”explains Nitin Chugh, country head – digitalbanking, HDFC Bank.

Eva is equipped to handle 70–80% of routine queries posed by HDFC customers. The bank’s website is visited by over 1.5 million users every month. “Eva is for people who like to chat and get information online… And she’s learning more (responses) on-the-job.,” Chugh adds. Eva is among 500 bots and virtual personal assistants (VPAs, like Apple’s Siri or Amazon’s Alexa) currently operational in India.


Over 90% of these are ‘chatbot’,which communicate with human beings using chat messages and are deployed on web portals and social media platforms such as Facebook. Most businesses use chatbots because it is easy to integrate on their communication platforms. Voice bots (speaking bots) may take more time to proliferate in India because of technological complexities and linguistic and dialectical differences.

Software research labs across the globe are locked in a race to create bots that are brainier than the rest - and mimic human beings the closest. These bots would, eventually, man the communication gateways of major private banks, telecom service providers, travel companies, e-commerce portals, private insurers and even drug manufacturers. “Asia is adopting bots much more than their peers in Europe and the US,” states Sindhu Joseph, CEO of bot-maker CogniCor Technologies, which serves companies like Tata Capital and Federal Bank.

Ram Menon, founder of US-based Avaamo Software, concurs: “Within Asia, China and India are the biggest buyers of chatbots. Almost half of our enquiries are from Indian companies,” he says. Proliferation of mobile phones and the tacit government-level push to go digital are encouraging Indian companies to include bots in their ranks. Beyond that, Indian companies are finding it difficult to set up more customer call centres, as these jack up operating costs.

“Indian companies cater to a lot of customers; their operational bases are huge. It makes business sense to employ bots to do routine, boring jobs… By doing so, they’ll be able to lower operational costs and also serve more customers at a time,” reasons Menon.

BOTTING UP
An internet robot, or simply a bot, is a software application that runs automated scripts on the web. Typically, bots perform tasks that are simple and structurally repetitive, at a much higher rate than what is possible for a human being. So, one single bot can handle queries (in whatever format – voice, chat or email) of hundreds of customers at the same time.


Automated scripts are responses plugged into stored (in mega servers or the cloud) knowledge banks periodically. The bot “reads” or “understands” the customer query and throws up an appropriate answer (from the stored knowledge bank). Over time, with constant manual “updation” of the knowledge bank, the bot is able to respond to even more complex customer queries.

Bots are cultured and customised to bring about a “human touch” to responses processed out. Many a time, this is done at the behest of companies that employ these bots.

“Companies don’t want their bots to sound too mechanical or business-like. So we do a lot of cultural adaptation in bots… over time, the bot learns local sensibilities and responds in an appropriate manner,” says Sindhu Joseph.

“We teach the bot a lot of colloquial terms to make the conversation human-like… We plug in words like achcha (OK in Hindi) to make the bot sound more like a human being,” she adds.

EARLY ADOPTERS
Almost all Indian private banks are at various stages of developing their own bots. HDFC, Kotak, Axis and ICICI Bank are running pilot programmes (involving higher level bots) across various platforms. “We are toying with the idea of enabling voice services in our bot,” says Amit Sethi, CIO of Axis Bank.

“Our chatbot will answer customer queries, explain product features, carry out transactions like funds transfer, bill pay and recharge,” he adds. Machine learning and artificial intelligence form the underlying base of bots. For businesses to start their bot programme, they should start collecting a lot of customer-related data.

This fragmented data is used to populate the knowledge bank servers. “Different types of data make the bot intelligent. Our attempt is to make bots that are intuitive and conversational,” says Deepak Sharma, chief digital officer, Kotak Mahindra Bank.

“Bots are also used within the bank – by our customer care executives… It helps our executives offer appropriate solutions to customers and also have a meaningful dialogue with them,” Sharma adds.

Indian banks, being the earliest of bot adopters, are on par with their global counterparts. Several Indian banks are now trying to feed bots with behavioural aspects of their customers as well. This will help banks to market their products better and also generate more sales leads. Bots, which understand behaviour pattern, will be able to pick “odd” transaction requests. This will go a long way in preventing online banking frauds, bankers opine. “There are chatbots out there which gauge customers’ mood from their typed language,” says Menon of Avaamo.

He explains further: “If you text a highly intelligent bank bot, {TRANSFER ` 50,000 TO SAM!!!} it will become conscious of your bold letters and exclamation marks; the bot will write back: {It seems you’re very angry today. Please speak to a customer care executive}. Such is the level of service offered by bots now,”Menon adds.

According to Sethi of Axis Bank, chatbots are already upping the customer experience. “Customers will no longer have to sift through irrelevant content on website before they find the right information or wait on call to speak to their phone banking officers,” he adds.

Bots have also made inroads into the Indian wealth management industry. “Robo advisory” services (investment advice) have become popular among Indian savers. FundsIndia. com, Scripbox and MyUniverse are some of the more popular robo advisory platforms in India. “Robo advisory is one area where bots have done exceedingly well,” says Tejinderpal Singh Miglani, CEO of Incedo, a bot-maker specialising in financial services, healthcare and telecom sector.

“Bots have helped users to take appropriate investment decisions. These programmes throw up investment calls after factoring in a lot of technical and fundamental parameters… and unlike human investment advisors, they’re not swayed by emotions, bias or greed for more commission,” believes Miglani.

Financial services institutions apart, telecom companies are also taking baby-steps to adopting bots in their day to day operations. A humongous customer base is becoming a hindrance in the widespread adoption of bots by telecom majors.

Nonetheless, companies like Vodafone and Airtel are running trials to put their bots on various communication platforms.

“Industries which have more B-to-C interactions tend to invest more in the implementation of bots. So you’ll see a lot of e-commerce set-ups, travel firms and insurance companies employing bots to handle customers,” adds Miglani of Incedo.

THREAT TO HUMANS
There is an increasing fear of bots taking over the jobs of human beings – and if it ever comes to that, the thriving BPO industry (call centres and voice processes) in emerging markets would be the most impacted.

“Businesses employ bots to reduce their cost and protect margins… If bots are put to work, businesses can save a lot of labour cost,” asserts Kiran Datar, vice-president –international sales, Moxtra, a technology company.

“Business will use bots for lower level transactions only… At that level, bots may eat into a lot of mundane jobs done by human beings. The counterpoint here is, higher bot technology will give rise to more high-value jobs where humans would be required,” Datar says.

The Indian BPO industry, according to NASSCOM, offers direct employment to over 30 lakh people. Close to 1.5 lakh new employees – mostly graduates -are added to the total workforce every year. Ironically, several BPOs have started pairing in bots to “assist” their (human) executives complete their jobs more efficiently.

“Intelligent enterprise chatbots can handle basic needs and help consumers perform transactions, but there will still be a need for human agents,” affirms Scott Horn, CMO at 24/7, an international software service company with BPO operations in key markets.

“Chatbots can allow organisations to work more efficiently by minimising repetitive work for agents. Used properly, they can reduce wait times by automated ‘commonly asked queries,’ clarifies Horn, adding, “They can even reduce the average handle time by suggesting responses while the agent is speaking with the customer, leading to faster resolution.”

Microsoft co-founder Bill Gates, in an interview with Quartz, opined that bots which takes away human jobs should be taxed on the income generated. Not many concur with Gates’ view though. “Even if such taxation existed, it would be difficult to manage and regulate,” says Horn of 24/7. Sharma of Kotak Mahindra Bank adds: “Bots will not replace humans in the foreseeable future. And when they do, humans will have enough high-value, high-paying jobs.”