Noshir Kaka, senior partner and global leader, analytics, and Nicolaus Henke, senior partner and global co-leader, digital and analytics at McKinsey & Company
Noshir Kaka, senior partner and global leader, analytics, and Nicolaus Henke, senior partner and global co-leader, digital and analytics at McKinsey & Company, are two of the most influential voices on technology in the world. Henke is also the chairman at advanced-analytics firm QuantumBlack. In an interview with , the technology leaders said that India was not only on its way to becoming an AI hub, it also has the unique advantage of being a generator of digital data, which is vital for AI algorithms to continuously learn and improve. Edited excerpts from the interview:
What is the potential of AI applications that you see across industries in the coming years?
Henke: If we apply AI to all business problems, the total value will be around $10-$15 trillion, while the world economy is worth around $80 trillion. The two dominant pools of usage will be in consumer industries like marketing, personalisation and promotions, followed by industrial and B2B companies
in manufacturing 4.0 and supply chain opportunities.
Why is AI so much in the news today?
Henke: The idea of AI was born in the 1950s. In the 1980s, machine learning
(ML) came into being, followed by its subset, ‘deep learning’ (DL), in the 2000s. The reason AI is so much in the news now is because 90 per cent of the world’s data have been created in the past two years and now we have the capacity to analyse that data using algorithms. Also, thanks to cloud computing, the cost of linking this data has reduced. AI makes forecasting cheaper and better using ML and DL techniques.
Kaka: Ajay Agrawal (eminent AI researcher and professor at Rotman School of Management, Toronto University) says that just as the internet reduced the cost of communication so we could start using it in ways not imagined earlier, we will also see the use of AI and ML-based models on all sorts of things when the cost of prediction goes down. If you are still using manual techniques for any type of forecast across sales, consumer behaviour etc, it’s almost a professional malpractice now. Historically, companies
have hired based on the order expectation for the quarter. Today, we can run AI on the order book and existing projects to predict who you need to hire to maintain your employee pyramid shape.
Where does India stand in the global AI landscape?
Kaka: We (India) have some of the largest databases and tech stacks on the planet, as in case of the UPI stack. One of the core uses of AI is in analysing data. We have access to both data and computing power. So the country has all the ingredients for becoming the AI talent factory of the world. And there’s no reason why India cannot leapfrog and come to the forefront in AI. Already, companies
across sectors are doing good work leveraging AI.
Countries like the US, China or Canada have invested heavily in AI research and development. Do you see enough support coming from the Indian government?
Henke: Governments can provide better capacity to train people in terms of skilling as well as using this technology as an enabler. But most countries have grown organically in this space.
Kaka: The government is already working as an enabler. But when it comes to innovations around AI, the native digital companies have done some remarkable work. For example, what Flipkart has done with its supply chain and logistics automation
is significant. Also, large banks have developed the digital front-end interface for customers at a scale that is a first in the world. We are not creating business use cases for automated driving because the business opportunity doesn’t exist yet. But when it comes to access in rural areas, we are innovating rapidly. We should celebrate the work being done in our own backyard instead of seeking inspiration from unicorns in other countries.
Globally, companies like the Carlyle Group and Overstock believe that AI is a market risk. What impact does McKinsey foresee?
Henke: Every technology has unintended consequences. For example, deep learning is changing the way call centres operate by enabling automation
through emotionally intelligent robots. However, the transformation is quite uneven. A large number of call centre roles might get automated and, hence, eliminated even though newer roles are not being created at the same rate. Moreover, there is a concern around bias in the data used to train these algorithms. If we start supporting doctors with machines, we need to be sure that the recommendations are safe. Then there are high-risk areas such as the deployment of AI in defence.
So, there will be an impact on jobs. But will a lot of new, AI-specific management roles get created?
Kaka: Progressive companies are making AI learning the default for their management. It means that they should be able to understand the impact of AI and how the system works. We find that while companies have data scientists
or engineers, they lack ‘data translators’, or those who help business leaders translate their business problems into something that only the data science team can understand. For one data scientist, you need five to seven data engineers and around 10-15 data translators. This role doesn’t need advanced maths or statistic degrees. It’s better to retain the C-suite number as it is and bring in a bench of translators.
Henke: Most organisations already have roles like CAO (chief analytics officer, CDO (chief digital officer), or CTO, which are distinct from the classic CIO roles.