Businesses Go for AI When Stakes are High

1230
Advertisement

In a conversation with Paromik Chakraborty of Electronics For You, Vaidya Subramaniam S., head-product management, GreyOrange, speaks about use of artificial intelligence (AI) in today’s businesses, things to consider before AI implementation and challenges ahead


Vaidya Subramaniam S., head-product management, GreyOrange
Vaidya Subramaniam S., head-product management, GreyOrange

Q. What is the level of intelligence in today’s commercial AI offerings?

A. Whatever we see today is a narrow AI or a weaker AI. Right now, we do not have machines that can act like a human being in multiple circumstances. Rather, systems go after a specific situation based on rules and heuristics today. Narrow AI is the stepping stone towards going broader.

Q. How is AI benefitting the various sectors? Please share examples.

A. Automation and robotics have been around for a while, helping out the manufacturing industries. Recently, this sector has started to incorporate intelligence. Balancing different manufacturing lines can improve productivity by 10-40 per cent depending on the degree of imbalance identified. Together with underlying data analytics, AI is used to balance the production lines to redesign the production process or to predict and address bottlenecks in real time.

In medical industry, risky and intricate surgeries, where even the smallest error could be fatal, are being done with the help of machines and intelligence.

Finance companies like Visa, which approve millions of transactions and issue millions of credit cards in a day, apply rules and logics along with the knowledge or information, and implement these over a system to carry out the different complex functions involved in the process, such as applicant credibility analysis and filtering, transaction validation and so on. These are the circumstances where AI is playing a huge role in today’s market.

In social media we can take the example of Facebook, which needs to monitor and filter out explicit or objectionable content on their public platform—and the filtering has to be done for all the media uploaded by millions of people throughout the day. Performing this activity at such a huge scale is not feasible without AI.

Q. What are the driving factors for implementing AI to a business?

A. Businesses go for higher technologies when the stakes are high and conventional methods are not sustainable or scalable. AI happens to be one of these technologies. Two major driving factors are the scale and complexity of the business.

Analytics is a low-hanging fruit for businesses of any level to catch upon. When a lot of capital is at stake, businesses will look at the available data, refer to the analytics and then take a decision—this creates a systematic way of running the business. Analytics is a precursor to AI.

Q. How can businesses scale up to implement AI and achieve success?

A. First, identify problems for which the business is not getting satisfactory results through normal means. A major reason is usually the ways of making business decisions. Next, figure out if there is a way to implement intelligence based on the past experience or statistical data and so on. Third, find out the areas and amount of expenses incurred and the factors involved. All these steps give you a clear picture of your investments and outcomes and helps you decide the technology necessary for attaining a success rate.

Q. AI itself is a complex subject. How can Indian companies harvest skill-set for the purpose?

A. The most complex aspect of AI is the self-learning capability—how to organise the knowledge automatically without any manual intervention, and how to learn adaptively. An easy point to start is developing an expert system or a room-based engine. To understand, assume a room full of experts distilling their observation and sharing their experiences of solving real-time problematic situations. If these experiences can be encoded in an automated manner, and if a computer follows that, you have your AI.

Other aspects of AI like operations, research, statistical techniques and probabilities are utilised to determine an outcome.


 

Advertisement


SHARE YOUR THOUGHTS & COMMENTS

Please enter your comment!
Please enter your name here