Ability to harness data without negating human intelligence to provide winning edge in use of AI: IBM executiveOctober 7, 2020
He noted that several sectors across retail, finance, healthcare and e-commerce are already deploying AI to increase efficiency and productivity, enhance customer service, and build a stronger security cover.
“These digital transformation journeys have further accelerated in the wake of the COVID-19 pandemic. As a result, the pace of digital acceleration compressed over a span of 8-10 months is something that would have actually taken years to happen,” he said.
The next step that will define the “winning edge”, he said, would be the “ability and efficacy to harness this data”, drawing descriptive and predictive analysis without negating the need and power of human intelligence.
“So, the important element and challenge is to harness the data, curate it and put it in context of enabling faster decision making. The data is the most powerful when we are able to apply it to our decision making at scale and speed, and to achieve this, we need AI at scale,” he added.
Patel noted that in order to capitalise AI to the fullest, it needs to be democratised, trusted and open.
“AI is only as good as the data that you’re using for it. And, as long as there is a lot of work which is involved in tagging and classifying the data correctly…companies have to trust it knowing that even after using AI, they will have full ownership and protection of their data and insights,” he said.
Patel added that AI must be fair (free from bias), accountable, explainable, secure and that openness is really important and key to driving continuous innovation.
The session also saw participation of Shalini Warrier, executive director, chief operating officer and business head (retail) at Federal Bank; and Tarun Aggarwal, assistant vice-president (engineering) at Maruti Suzuki.
They cited examples of how AI is being used at their organisations to drive efficiency and personalisation of services to enhance customer experience and engagement.