New Delhi - Among 13 global industry sectors, the insurance industry is leading expenditure in Artificial Intelligence (AI) systems with $124 million, a study by IT major Tata Consultancy Services (TCS) said.
According to the study which covered North America, Europe, Asia-Pacific and Latin America, the average expenditure on AI systems was $70 million across the sectors surveyed.
Next to insurance, consumer packaged goods reported the second most significant expenditure with $95 million.
"It's striking that the sectors making the boldest AI investments seem to group around industries like insurance, travel, hospitality, telecom and consumer-focused sectors such as the consumer packaged goods industry," said Ananth Krishnan, Chief Technology Officer, TCS, in a statement.
The study said that 80 per cent of the executives in all 13 industries have currently invested in AI and almost 100% plan to invest by 2020.
The increase is sharper among industries with lowest current investment levels, including the travel, transport and hospitality sector, which plans to increase its spend by 750%, from $4 million in 2015 to $34 million to 2020.
It is followed by rises in media, entertainment and information services (292%), industrial manufacturing (74%), healthcare (44%) and banking and finance (29%).
In terms of percentage of average company revenue, the consumer packaged goods (CPG) industry led with 0.66%, followed by utilities (0.53%), insurance (0.52%) and telecommunications (0.39%).
The most frequently mentioned business function that used AI was IT. Apart from IT, only 29% of companies are using AI in sales.
However, more than half of the CPG companies (52%) and nearly half the retailers (49%) are using AI tools to improve sales performance.
Earlier, the first phase of the study underlined that in almost all sectors, addressing people's fears about losing their jobs was not ranked as a major barrier.
The companies with biggest revenue and cost improvements from AI see the need for at least three times as many new jobs in each function by 2020 because of cognitive computing innovations.