Artificial intelligence has long been recognized as one of the most disruptive technologies on the market, and going by the research, it will continue to be a leading ‘game changer’ technology this year. A report by NTT found that AI underpins the top five data-driven disruptive technologies in 2020; similarly, six out of Gartner’s top 10 picks for strategic technology trends this year are heavily powered by AI.
In other words, companies that do not jump on the AI bandwagon will most likely be left in the dust. But that doesn’t mean everyone should start using AI just for its own sake. In order to get real value out of the technology, companies need to exercise some thought, some effort, and some investment that goes beyond the merely monetary.
Before you begin, understand what you plan to do with AI
Like any other business decision, implementing AI must be done with a specific objective in mind, said Tong Hsien-Hui, head of venture investing at deep tech startup incubator SGInnovate. Speaking at a recent conference on tech trends and predictions for 2020, he explained that the chances of success are higher when people actually identify a bottleneck first and then use AI to deal with it. “Companies that succeed look at a specific use case,” he advised.
Other speakers at the conference, which is organized annually by law firm RHTLaw Taylor Wessing, similarly cautioned against using AI just because it appears to be the latest trend. “Don’t just implement AI because your competitor is using it, or because it’s a new toy!” said Richard Wong, head of ICT at Frost & Sullivan. “Do we really know what is the problem we want to solve using AI, before we use it?”
Suresh Shankar, the founder of big data firm Crayon Data, said that companies need to understand the difference between a solution in search of a problem and a problem in search of a solution. “Unfortunately, most companies are doing the first [implementing a solution despite not having a problem that needs solving],” he said. “The drive to implement AI has to come from a leader who has a problem, and has a clear idea that conventional solutions are not going to work to solve that problem, and needs to go to the next level of technology.”
Don’t try to solve the problem by yourself
Businesses that expand internationally are typically advised to work with partners who can ease their path in the unfamiliar waters. The same goes for AI and other unfamiliar technologies, say the experts: and it’s not just because they need the help. It can be incredibly difficult for a non-AI-focused to put together a dedicated AI team, simply because professionals are not interested in the work.
“If you are a person who is very skilled in AI, you won’t want to work in a company that is just working on one problem. You will want to work somewhere you can work on many different problems,” said Shankar. Hence, instead of trying to hire your own AI professionals, he recommends outsourcing the work to a company that is already experienced in the field.
Perhaps because AI has proven so challenging, businesses appear to be more conservative about their ambitions this year. PwC’s latest AI Predictions survey found that only four percent of executives plan to deploy AI organization-wide this year, a drastic drop from 20 percent last year. The change is not explained away by successful implementations, either: only 18 percent have implemented it in multiple areas, down from 27 percent in 2019. Instead, the majority (42 percent) are simply investigating its use this year, followed by 23 percent carrying out pilots in discrete areas.
Upskill your people and start early
Success in AI is usually found by companies that are already digitized, with structured systems and steady data inflow in place, said Tong. A great many SMEs, however, lack that, and these are the businesses most in danger. “It’s one thing for the [boss] to say, ‘I’ll forgo my new Mercedes this year and we’ll get an AI system’, but the company may not be ready,” he observed.
The solution is to start training people, and do it soon. There is already an appetite for AI upskilling: Udemy’s 2020 Workplace Learning Trends report found that over the last three years, trending skills have shifted towards AI and data science, and predicts that AI skills will again top the list in 2020. Getting the workforce ready to use AI will lay the foundation for the next step, which is to actually start working with data.
Moreover, companies should not be concerned about having very little data, or not having it digitised, said Shankar. Getting started is the important thing. “You may have less data, but once you start to use it, you can very quickly catch up with a company that has more data [but did not start as early],” he pointed out.
And what would success look like?
If AI is properly treated as a business solution, rather than a magical catch-all panacea, success is actually fairly straightforward to measure, according to venture capitalist Tong. Companies can look at easily collected metrics such as lower cost, better services, or increased revenue.
More importantly, though, success may be seen in terms of how future-ready the entire organization is. Statistics from Frost & Sullivan indicate that the penetration rate of AI is expected to more than double in the next six or seven years, meaning that companies which don’t start preparing their workforce and business processes very soon will find themselves falling far behind the competition.
“As human beings, we know AI is coming,” said Wong. “What are we doing in terms of upgrading our skills, to ensure that we are able to use it?”