Artificial intelligence (AI) can generate X-ray reports at Max Healthcare. This evolving tool — in beta stage at present — is being programmed with radiologists’ knowhow, by scanning millions of patient records.
And though it may not yet be able to read the blur on the film as tuberculosis, in five years, AI could make that interpretation as well.
On PolicyBazaar.com, almost 70% of motor insurance policies are sold by bots. HDFC Bank products such as ‘loan in 10 seconds’ are processed by machines.
At travel platform Ixigo, bots handle most queries without people realising. Machines are adept at checking chances of a waitlisted ticket being confirmed, lowest fares, the best time to book tickets and so on. There’s no human involved at any stage.
Apollo Munich, Birla Sun Life, ICICI Lombard GIC, Maruti, Yes Bank and many others are actively using AI in daily business — improving productivity and scale. Ritesh Pai, chief digital officer, Yes Bank, sees AI platforms do voluminous repetitive tasks and improve 245;7 functioning by servicing even NRI customers, in any times zone. “AI is an opportunity companies can use to their advantage,” says Pai.
Getting personal again
AI, or the ability of machines to mimic the human mind, is slowly but surely invading corporate workplaces. Arup Roy, research director, Gartner, says, “Machines are enabling intelligent conversations. Banks, insurance companies, wealth management firms, e-commerce and retail firms are extensively using machines for predictive analysis, sales and so on.”
Healthcare AI can diagnose, assist with medication reminders and even develop treatment pathways for diabetes or cancer. In not so distant future, “machines could ease the pain for doctors to sift through past records (which machines will do) and focus more on critical healthcare needs,” says Rohit Kapoor chief growth officer, Max Healthcare.
PolicyBazaar expects AI will be good to sell 85% of motor insurance policies over the next few months. It will soon start automating the sale of health, term and other policies. Ashish Gupta, chief technology officer, PolicyBazaar, says, “AI is transforming the way we work.” The motivation is not just to keep customer acquisition costs low but also ensure that if the business grows 100%, manpower expands merely 10-15%.
HDFC Bank machines have journeyed from assisting humans by sending alerts to taking over repetitive processes such as monthly account balance updates or chequebook requests. Today, they can check on loan eligibility or create a risk profile of a customer. Processing loans, which took at least two months earlier, is now done in seconds, minutes or just a couple of days.
Nitin Chugh, country head, digital banking, HDFC Bank, says, “Banking used to be a very personal experience. Digital made it impersonal. But now with AI and customisation, it’s becoming personal again.”
The bank’s AI chatbot, Eva, has already handled more than five million queries with 85% or higher accuracy. Analysts predict that by 2020, AI technology will be virtually pervasive in almost every new software product and service. Gartner predicts AI to be among top five investment priorities for more than 30% of chief information officers by then.
However, there are certain limitations to a machine that is not human, neither a clone, but a very clever software. It can mimic basic functions and is good at repetitive tasks, but need a lot more time to fathom the complexity of human interactions. For instance, at Max hospitals which is beta testing machine applications, AI algorithms are tested by radiologists and tweaked if need be to improve their knowledge.
Even if the AI tool is as accurate as a human, hospitals don’t want to take chances and have both humans and machines working together. Though this could change ushering in a leap for AI from labs to reality. Machines have comprehension issues with typographical errors, colloquial interactions and languages other than English or Hindi. The PolicyBazaar AI tool — based on Google platform API.ai and rebranded as Dialogflow — works well in English and Hindi but not regional languages.
The insurance platform has now started training it in Tamil, Telugu and Malayalam. Besides, Gupta does not see AI assisting beyond 85% of PolicyBazaar’s sales. “We need natural language processing, which may not be good enough to close all sales call,” he adds. If you want a loan in 10 seconds, there’s no room to negotiate. To take care of nuances such as a lower interest rate or deferred payment plan, a human will have to step in.
Nonetheless, “The future needs to keep pace with technology,” says Chugh of HDFC. In tasks where a human does not add value to the company, the machine is going to be its most valuable asset. It improves productivity, cuts costs and helps deploy humans in more ‘thinking’ roles. Kumar of Ixigo believes we are already in an era of hyper-personalised experiences. So while it may be artificial, it’s still intelligence.
Will Artificial Intelligence Take My Job?
Technology versus employment is an age-old debate. But it’s becoming more and more real with each passing year, as companies lean towards automation to improve productivity and margins. Consider the numbers. A bank recently shrunk its card reconciliation team from 200 to just five. A 500MW thermal power unit manages with just 200 people today, compared with 2,000 a decade ago. PolicyBazaar plans to double its business without adding much headcount.
Thanks to AI, bots, and algorithms, jobs are disappearing at an ostensibly faster pace than they are being created. K Sudarshan, managing partner, EMA Partners International, a C-level headhunter, says, “Operations across companies will be impacted by the use of AI.” A corporate honcho who wished not to be named quips, “I would not want to be in college at this time.”
However, Manish Sabharwal, chairman of TeamLease Services, the largest temp staffing company in India, feels, “AI might be relevant at the company level but not at country level.” He points to India’s low per capita income of about $1,650, where World Bank ranked India 112th of 164 countries.
Countries with higher per capita income (such as US with $45,000, Singapore with $50,000 or UK at $37,000) will use more AI platforms than India, as their labour costs are very high. In fact, Sabharwal sees technology workforce doubling to 8 million in just five years as more companies adopt technology tools.
Ritesh Pai, chief digital officer, YES Bank, says, “When video conferencing started (back in the mid-1990s), people thought corporate travel will reduce, but that has actually grown 20% on-year. AI will assist humans, while humans will step in whenever complexity increases.”
AI doesn’t just take over jobs; it creates new ones. For instance, a German automaker in India has a 15,000 team with 1,000 engaged in making cars — the rest in coding. In India, manufacturing employment may peak at 17-20% of the total jobs — up from 11% at present — while most new jobs will be in the services sector. It may well end up being a question of reskilling, rather than losing your job to AI.