Andre Howson
On any list describing the technological trends of the future, you’ll find AI. In fact, as Kevin Kelly, the founding executive editor of Wired, proclaims in his book, the proliferation of AI is ‘inevitable’. Artificial Intelligence is going to shape everything—from the way we drive our cars, to the way we do business, to how we flip our burgers. The very idea of AI has captured the imagination of the entire tech community. From fears that humanity is developing the next Skynet to perhaps more legitimate worries that we are biting off more than we could chew, AI has one foot in the future and the other firmly in our present collective consciousness.
Which is not always a good thing. As with anything that gets popular in this hyper capitalist world, the term can be misused.
Practically everything is being marketed as AI, even when it isn’t, and the hype around the word can be leveraged for unwholesome purposes.
Despite its growing pains, we do know AI is going to be a very big part of industry in the not too distant future. If we want to ride the technological curve, Asia needs to get its machine-learning-surf-boards and start paddling before we miss the wave.
Can we compete with countries who are so much more technologically ahead—and so much better funded—than we are? Well, yes. Because AI, as it is right now, is not all that fancy, and coding is just the “next big blue collar job”. And who does the best at ‘stealing’ blue collar jobs? That’s right. Asia.
While leaving the more exciting applications of AI to Google and Elon Musk, Sri Lanka can get on the more ‘mundane’ and perhaps the more scalable aspects of AI solutions.
The Artificial Intelligence Industry In Sri Lanka
Sri Lanka does have its own Artificial Intelligence scene going on. The University of Moratuwa offers a Masters degree in AI. There’s also a Sri Lanka Association for Artificial Intelligence (SLAAI). We didn’t link to their website because it is infected with malware, which is sort of ominous. The association, however, does conduct regular meetups and lectures on AI. But as far as commercial industry goes, there’s really not that much happening. Public Awareness and understanding tends to be on the lower end of the spectrum.
It’s not that there is nothing commercial happening with AI in Sri Lanka. Earlier this year, the Colombo Centre for Cognitive Computing (CogCom) opened its new office in Orion City. Now CogCom is not a solo entity but a subsidiary of Volume Ltd., a UK-based global provider of web-interaction applications. What’s interesting about Volume is that their services are powered by Artificial Intelligence and Machine Learning—a sort of high-powered customer service and interaction platform.
Volume works with several high-profile brands to power and automate certain services. Their clientele includes Cisco, Castrol, Dell, Epson, Oracle and Virgin Media Business.
For Virgin Media Business, for example, they created a sales performance application to train their sales teams to perform better. They also have digital concierge applications and technical advisor applications, as well as employee well-being applications that use a variety of sources to digitally monitor employee well being—a sort of virtual buddy. But probably the best example of what they do is Volume’s own website, which is sort of atypical. Instead of clicking through navigation links, you type your question on to a text prompt and have the answers read out to you by Lusy. Lusy is the voice and the programme that powers the entire website. It was created using Watson, a cognitive system developed by IBM. Watson is the reason Lusy is able to answer your questions using natural language. Its domain is pretty much limited to all things Volume, so there’s no point asking it the meaning of life (yes, we tried).
Lusy is the domain specific AI behind the Volume website
CogCom is Volume’s branch in Sri Lanka and is very much involved in Volume’s global work. We spoke to Asiri Fernando, the General Manager of CogCom to find out what the state of AI is in their particular industry and to find out what the future may hold.
AI Examples
Part of the reason so much of Volume’s and CogCom’s work involves creating digital assistants and automating learning processes is because these are the assembly lines of the digital world—processes that are repetitive and take up so much time and manpower. According to Asiri, companies come to them when their technical experts are tied up answering the same sort of questions from customers over and over again—which is around when they decided to automate the process using artificial intelligence solutions.
With Castrol, the automotive lubricant company, what they did with AI was to automate the first point of interaction with customers. The AI would answer all the low-value, repetitive questions, leaving the technical experts to deal with the higher-value, more complex issues personally.
With Glooo, what they built was a well-being monitoring application—sort of like a virtual buddy. The application was created to deal with the global mental health epidemic. It builds a relationship with the employee over time, analysing sentiments and adapting to the employee’s current emotional state. Glooo can alert HR to sustained low moods or out-of-character behaviour. It can also show empathy and put employees in touch with someone who can help them.
The Digital Concierge application is what Lusy is. A web application that gets better the more it is used, and one that replaces the traditional drudgery of hunting for content through a web page.
Applying AI In Sri Lanka
All these types of solutions are available in Sri Lanka as well, but while businesses are excited about the whole concept of AI, they don’t know what to do with it. According to Asiri, the telecom sector is where AI applications can shine best. Telcos host a lot of data and that is where AI can be most useful, sorting through this huge amount of data and automating high-volume processes and human interactions like customer service.
Customer service in Sri Lanka can be a difficult thing, with a high turnover of staff and a lot of time spent training them. An AI application can automate the process on multiple levels, handling low-level customer queries and fast-tracking the training processes of operators. AI can also set the tone for interactions with a customer, possessing either a friendly tone of voice or something more formal, depending on what a brand needs.
Volume has experimented with combining AI and robots, as a means of creating meaningful interactions in customer service. But, as always, solutions need to be made with context in mind. For example, at one particular high-end car showroom, the company proposed a robot/AI solution that would greet customers, collect their information and then pass on the information to a test driver when the time came. But the customers who typically visited the showroom are on the higher end of the income bracket and preferred dealing directly with other humans. However, in another test case, when the solution was proposed at a showroom that sold cars that were more family oriented, the children and accompanying family members would spend time interacting with the robot while the buyer goes over the car.
Volume used robots from Softbank, the Japanese robotics manufacturer, in some of their solutions. Image credit: Japan Times
AI has many uses, but as always the context in which it’s placed is important.
The Challenges
Asiri says that part of the challenge in implementing AI is training the application. In the beginning, the application is like a child. It needs to be trained using large amounts of data, then the model needs to be tested to see how accurate it is. If there are errors, adjustments have to be made and the data needs to be refined till the model works.
Asiri believes the knowledge for this is pretty much available in Sri Lanka, but that it just needs to be applied to business. With the proliferation of big data and the popularity of data science careers, AI has a solid base to build on. Sri Lanka could very well grab a corner of the AI market.
We don’t even need to look to the West to sell our product. According to Asiri, even European markets have not fully adjusted to using AI as part of their businesses. But he states that by 2018, at least 20% of companies are predicted to budget for AI. So there will be plenty of opportunities in the Asia Pacific region for AI solutions.
He suggests that the transition to AI will happen very smoothly. Just like mobile adoption increased in Sri Lanka and how cloud platforms became part of most businesses, AI too will be inside our systems before we are even aware of it.
Most of the industry sectors that AI will affect will be the unglamourous sort—process automation, customer service and data crunching. Perhaps they are not as exciting as talking computers and self driving cars, but this is where the core of commerce lies, and this is perhaps the best part of the market for Sri Lanka to corner.
