Laurent - stock.adobe.com

India ramps up on AI amid talent and scalability challenges

Indian organisations are speeding up deployments of AI across multiple sectors, but legacy systems, siloed data and a shortage of AI-specific talent will stand in the way of greater adoption

This article can also be found in the Premium Editorial Download: CW Asia-Pacific: CW APAC: Trend Watch: Artificial intelligence in APAC

At one of Miraj Cinemas’ movie theatres across India, an artificial intelligence (AI) algorithm is at work, counting the number of occupied seats in real time and feeding that data into a dashboard for the cinema operator to track occupancy rates.

Dhawal Mehta, general manager for systems and IT at Miraj Cinemas, said the automated occupancy detection system, custom-built for cinema halls, is non-invasive and can reduce the need for labour-intensive visual inspections.

“We now have real-time data, enabling us to compare historical occupancy data across different spaces to drive insights on patron behaviour,” Mehta said, adding that the insights also help to improve the confidence of movie distributors in Miraj’s patron occupancy rates.

If cinema seats are finding new eyes, then factories are not far behind. Hindalco, a supplier of aluminium and copper materials with 15 sites and thousands of assets, has turned to AI to fix problems with its assets before they fail.

By training machine learning models to predict anomalies and failures, Hindalco’s predictive maintenance initiative has helped the company save hours of downtime at one of its plants and boosted annual production by 110 tonnes.

Biswajeet Mahapatra, principal analyst at Forrester, noted that the adoption of AI in India has been moving in the right direction, adding that the focus has shifted from generating business insights to automating business infrastructure.

“We have also seen AI and machine learning being used for digital twins that span automation, real-time decision-making, and strategic modelling in India,” he added.

One reason for the growing use of AI in India has been the ongoing efforts by the government to shore up the country’s manufacturing capabilities.

“Due to the Modi government’s Made in India focus, manufacturing has come to the centre stage as much as services,” said Rohit Kochar, founder, executive chairman and CEO of Bert Labs, a technology service provider that specialises in AI and industrial internet of things (IoT) applications

“For India to compete with other manufacturing hubs like Southeast Asia, India had to adopt Industry 4.0, and by extension, AI applications,” he added.

We will be lagging behind as a provider of talent to the world unless we develop rounded professionals who can think of AI in a true sense
Jitendra Singh, TalentSprint

That said, India’s strong services sector has been a cradle for AI deployments too. “As India is home to many back offices, AI adoption is predominant in internal process automation and in driving revenue growth,” said Mahapatra.

According to the AI game changers report 2022 by IndiaAI, a government-led industry group, AI is expected to raise India’s annual growth rate by 1.3% by 2035 – an addition of $957bn, or 15% of current GVA (gross value added) to India’s economy.

India’s dream of becoming a trillion-dollar digital economy could well be tied to the rate of AI adoption. A separate report by the National Association of Software and Service Companies (Nasscom) noted that four key sectors in India could contribute about 60% of AI’s potential value-add of about $500bn to India’s GDP by 2025.

As a sign of the growing momentum around AI in the subcontinent, 65% of AI prototype projects in India have reached production scale compared to 49% globally, according to a study by Bain & Company, Microsoft, and the Internet and Mobile Association of India.

Talent and scalability challenges

Indian organisations continue to be hamstrung by legacy systems and siloed data that stand in the way of higher AI adoption. Other challenges include low maturity levels of the country’s AI ecosystem and the absence of advanced metrics to quantify returns from AI investments.

Jitendra Singh, chief technology officer at TalentSprint, noted that the advantage – and challenge – for India when it comes to AI is the country’s large population.

“The solutions that work easily in the western world become very difficult to use in India because of our scale and complexity,” Singh said. “Given that we produce the largest number of software and computer science engineers in the world, we should be at the cutting edge of AI development.”

But while tech talent in India is plentiful, not all will make the cut when it comes to AI. That’s because many people view AI as just another technology like Android or Java, according to Singh.

“It’s a mistaken approach,” he said. “AI is not a toolkit or a set of commands to learn. One of the key gaps in learning AI is maths, which was almost never a requirement for a software engineer in real terms. We will be lagging behind as a provider of talent to the world unless we develop rounded professionals who can think of AI in a true sense.”

What’s different about India, as Bert Labs’ Kochar weighs in, is its ability to view AI as part of a vision, strategy and execution for the long term. That’s what separates India from countries like US, Germany, France, Italy and China, which tend to deploy AI more in industrial sectors – albeit at a faster rate, he said.

Work has started to plug the talent gap. Several institutions, such as the Indian Institute of Technology (IIT), have started undergraduate and graduate degree programmes in data science and AI, said Debanga Raj Neog, assistant professor at IIT Guwahati. “The goal is to provide a workforce ready to use AI for understanding and tackling challenging and unique problems faced by our nation of 1.3 billion people.”

Neeraj Kumar Sharma, another assistant professor at IIT Guwahati, said on brighter side, India can avoid errors in the initial deployments of new technologies, such as AI, and pursue developments in an efficient manner.

“Every now and then, India releases innovations which surprise the developed world. For example, putting 104 satellites into orbit in one launch, evacuating one million people on short notice before a storm hits the coast, operating ‘faceless, paperless and cashless’ services in banking and purchase transactions, and enabling cheap access to mobile internet,” he said.

Regulatory nudges to spur AI adoption by insurers

A slew of regulatory changes in India is set to pave the way for greater adoption of AI in the insurance industry. In July 2022, the Insurance Regulatory and Development Authority of India (IRDAI) said it would permit insurance firms to provide tech-enabled add-ons to motor insurance policies.

IRDAI chairman Debasish Panda had also encouraged insurance players to make health insurance products affordable, hinting at the untapped scope of AI and customisation.

Another major move was the relaxation of product approvals with the expansion of the Use and File process to include life products. Previously, life products needed regulatory approval before their launch, as per the prevailing File and Use policy.

Fresh from the oven are hackathons where IRDAI is welcoming ideas to develop technology-driven innovative solutions for automated death claim settlement in a bid to curtail mis-selling and enable technology-based distribution of insurance products in areas of poor penetration. These are major turning points, giving insurers permissions to conduct video-based know-your-customer (KYC) processes and launch standardised insurance products.

Even before the regulatory nudges were announced, Indian insurance companies have also dipped their toes into AI, some in a big way.

For example, ICICI Prudential Life Insurance is using AI-powered chatbots and assistants for renewal calling, and to sell pre-approved policies with customised solutions that enhance the buying experience. It is also using AI in areas like fraud prediction and risk assessment.

Others like Aegon Life Insurance have introduced algorithmic underwriting while Onsurity is working on a predictive deep learning model to identify its customers’ claim amounts.

Read more on Artificial intelligence, automation and robotics