How firms are using AI, automation, data analytics

India’s digital push holds the potential to transform the country into a data-rich economy. Industry body NASSCOM has pointed out that AI will play a crucial role in realizing the government’s vision to turn India into a $5 trillion economy by 2024.   

Many companies are increasingly gathering data, but only a small fraction of this data is analysed and made use of. Moreover, the insights will be as good as the data that is being fed.  

In a panel discussion moderated by Leslie D’Monte, Executive editor, Mint (TechCircle), at the Mint Digital Innovation Summit, industry experts shed light on how they are using AI, automation and data analytics, the varied and interesting use cases for automation, analytics and challenges that lie ahead for them in the industry.  Edited excerpts: 

On using AI and data in operations

Amod Malviya, Co-founder, Udaan: Data has been central to Udaan from the beginning. The problem we tried to solve is fragmentation in the B2B space. Nothing is standardized. We have to use every technology we get our hands on in order to get insights into the market and work with the same efficiency as a large centralized B2B player. The challenges have been the quality of data, in this regard, data governance has to be incorporated as a first-class concept. In Udaan we take data governance seriously. We look into data not as a by-product of the lifecycle but as a crucial differentiator.   

The importance of AI and data analytics for the telecom sector

Adarsh Nair, CEO Airtel Digital, Chief Product Officer, Airtel: Many parts of Airtel are deeply invested in data and data products. India has one of the lowest average revenue per unit (ARPU) per person, yet we are a billion people. This is a problem where technology has to come in to solve. 1000s of towers are deployed every month to get better coverage.   

We used to install towers earlier through people surveys and through physical estimations, which took us weeks to roll out towers. Today the methodology is powered by data analytics, AI and ML. We use maps of India, utilise computer vision to identify various parameters. Using this data, we identify where the ideal tower should be.   

 Today agent calls are filtered through a speech recognition engine that analyses every call and understands sentiment patterns such as anger to improve quality of training and quality of solving issues.   

Using AI in the BFSI sector

Deepak Sharma, Chief Digital Officer, Kotak Bank: In the first wave we learnt that converting human interactions into digital experiences was crucial. In the phase 1 the chatbots were working independently from sister channels. A lot of complex conversations were not something that these channels were used to. Customers needed answers to complex queries.  In the first wave we scaled our conversational AI from 80,000 skills to almost 200,000. We asked our employees who couldn’t make it to office to train the engines on what they know. That is how we could internally crowdsource a lot of interactions from human to AI intelligence.   

Using AI/ML in the area of Microfinance

Suresh Shan, Head of Innovation and Future of Technologies, Mahindra Finance: When you try to predict something in rural India, you have to address specifically the connectivity, electricity and resource availability in the last mile. The plan is to keep the technology simple and utilise as much of the data as possible. You can see a big shift from corporate digital to customer digital.  More technologies and services are now focused on the customer. The model has evolved from collecting information to just process a loan, to creating meaningful engagements with the customer. You have to be aware of how the demography, disparity and emerging needs/ current needs are taken care of.   

Automation and AI go hand-in-hand

Nitin Purwar- Industry Practice Director- APAC & EMEA, UiPath: We have evolved from pure-play, back-office, rule-based automation to a lot of cognitive automation. There are a few reasons why this has happened, one is the rising importance of data. Secondly, the convergence of different technologies, for example, automation today includes many machine learning, RPA, document processing, API based automation. When we look at customer experience, it is across channels. We have done lots of contact center automations where customer experience can be improved by optimising time.  Today, the number of emails being sent by enterprises has increased multi-fold, which has given rise to smart-email automation. Automation is more about enabling an omni-channel presence to provide a better customer experience. Covid also saw the rise of many prediction capabilities using machine learning models.   

On terms like AI natives and AI migrants?

Jayant Kolla, Partner, Convergence Catalyst: When we talk about AI natives and migrants, we are referring to AI maturity and adoption. Globally companies can be segregated into 4 types based on AI maturity- 20% form the beginners, 32% are implementers, 33% are advancers while 15% are leaders. Right from investments, AI use cases, ROI, quality of data varies according to the AI maturity. Not many companies are addressing the elephant in the room in terms of use-cases in AI, which is the Covid Pandemic. 20-21 has changed the problem statements for AI, automation and data analytics.