Say you have a complicated task at hand, like reading the next paragraph of this article in a language you don’t know. That’s certainly daunting, but you can solve it easily with the help of an expert who does know the language, thanks to clever technology that exists today. Welcome to the world of artificial intelligence (AI), machine learning and cloud computing.
If you think about it, these three elements exist all around us all the time. This is, after all, a website, and those are all on the Web, which is facilitated by the Internet. Today, cloud computing powers the Internet, while also needing the Internet to function. Indeed, we take it for granted and need not know the first thing about how it works to make it work for us. Given that this is a listicle — as they call it in the trade — we have to recognise the Internet as the original cloud computing surprise so we’ll call this point number zero. So the question then is: “How can you better deploy AI, machine learning and cloud computing to meet your business needs?”
1. Machines are Already Making Decisions
It is important to bear in mind that we are only looking at what’s called Narrow AI, not General AI, which is (still) the stuff of sci-fi. Narrow AI is phenomenally good at using data to make decisions, in a very specific way. Making decisions is the key and exciting part in this equation, because machines don’t typically do this. IBM Watson famously won a game of Jeopardy way back in 2011 against the best human competitors with its wealth of knowledge on topics ranging from ancient languages to fashion design.
2. Companies are Using Watson to Power Chatbots
With the right amount of training, Watson can also be used to power chatbots just as how insurance giant Prudential has done. IBM used Watson to build askPRU, Prudential's cognitive-powered chatbot. Integrated into Prudential’s back-end systems, the chatbot is able to instantly retrieve data such as a customer’s policy cash value, policy premium due date and status of submitted claims, among others. The 24/7 chatbot was trained by data scientists to understand non-scripted questions, and deliver responses a human would easily understand.
Obviously, this goes far beyond merely providing answers to FAQs on demand, and it isn’t about taking human beings out of the equation. In fact, it’s quite the opposite, as Ms. Theresa Nai, Chief Operating Officer at Prudential Singapore, explains:
Providing our customers with a seamless experience is important to us. With the askPRU chatbot, our Financial Consultants have round-the-clock and instant access to customer-specific policy information which allows them to respond to their customers’ queries even more promptly. This is one of the ways in which we digitally enable our Financial Consultants so they are equipped to meet the increasing consumer demands in today’s fast-paced world.
3. Artificial Intelligence Needs Human Trainers
And if you’re worried that machines are going to get so smart one day that they end up enslaving us, the truth is less dramatic: AI can only do exactly what you train it to do. IBM Asia-Pacific Chief Technology Officer for Analytics Kitman Cheung explains the general idea this way: “What we’re doing (in general, from hardware to software) is expanding the sensory network of AI, and then using machine learning and data science to formulate the best course of action from (all the accumulated) data. Finally, a human being uses this advice to make the best decision possible.”
Of course, training a machine like Watson to be useful to humans is challenging but the consensus today is that this process must be as simple as installing Windows on one’s own home computer, for example. IBM Watson’s Cognitive Solutions Architect Manprit Singh explained: “We make the whole AI infrastructure fully trainable, and you only need domain experts to do this. It was important to us (from the start) that you don’t need mathematicians and PhD holders to do this because (ease of use) is key.”
4. Machine Learning Will Augment Human Expertise
“It (Watson) may be overkill (as the force behind a chatbot) but what we want to do is elevate the workforce.” Machine learning — including everything from analytics to AI — is best when it augments a human being’s productivity.
In the example of askPru, Prudential Singapore’s call centre has seen a 30 per cent call reduction since the chatbot’s launch. This is due to askPRU’s ability to gauge a user’s intent and to accurately respond to a diverse set of common queries, freeing up time for the firm’s call consultants to handle more complex questions and deliver higher value work.
Discover more about the exciting prospects that AI presents for your business, or get up close with these innovations at Think ASEAN — a gathering for the innovative, curious and persistent. Find out more here.