CEOs are struggling to take AI from experiment through to implementation. It shouldn’t be surprising, then, to find that certain sectors have pushed ahead of the curve.
It isn’t uncommon for publications to position an innovation as “the next big thing” by comparing it to the Industrial Revolution or the Internet. After all, nothing sells quite like hyperbole. Perhaps that’s why 47 percent of CEOs thought that Artificial Intelligence was over-hyped just four years ago, according to a PwC survey.
That being said, even a cautious observer has to admit that AI has cut its teeth in the years since. From information technology to pharmaceuticals, industry leaders have experimented with a wide range of cognitive computing systems such as AI.
These technologies are called cognitive because they “understand, reason, learn and interact by continually building knowledge, understanding natural language, and reasoning and interacting more naturally with human beings than traditional programmable systems.” Though these capacities make AI systems a puzzle for philosophers, they open a new frontier for the real-world application of cognitive computing.
Today, a wide range of experiments with cognitive systems are being scaled to enhance daily operations. Yet change isn’t visiting each industry at the same time, nor progressing at the same pace. Many of the differences between sectors are driven by underlying opportunities–customer-facing telecom providers, for example, will find a greater scope of application for AI than Infrastructure organizations. However, the companies that are making the greatest use of AI have adopted a suite of policies that are far from common, indicating that AI optimization is highly dependent on the strategic decisions of CEOs.
Differential AI Adoption
Even though AI is a new technology, huge differences in adoption rates across industries have already appeared. Yet there is no telling whether these differences will survive the next decade.
Surveys show that companies see an opportunity to employ AI to enhance IT, information security, innovation, customer service, and risk. These goals may be industry-agnostic, but in practice, there have been both early and late adopters. As of 2019, the three industries at the top of McKinsey & Company’s AI capability ratings were high tech, automotive and assembly, and telecom, respectively. Yet they were far from the fastest-growing sectors for AI. From 2018 to 2019, the number of retail organizations with at least one AI capability embedded jumped by a whopping 35 percent. Travel, transport, and logistics weren’t too far behind, with a 26 percent increase over those 12 months. If these surveys are any indication, it appears that the field could look very different in just five years.
Of course, AI already looks different from one industry to the next. Whereas high tech is heavily reliant on machine learning, automotive companies embed robotic process automation, as well as physical robots, into operations on the factory floor. Some companies have even used AI to launch into new industries.
Toshiba Electronics Taiwan Corp. is a case-in-point. This consumer electronics company recently developed a wearable AI device equipped with biometric sensors that collect a constant stream of data, including heart rate and blood oxygen. The device then interprets patterns to identify unhealthy trends and alert caregivers. By offloading this crucial diagnostic task to the device, healthcare professionals can concentrate their expertise on cases requiring a human response.
Retail, on the other hand, sees the widespread implementation of natural language text understanding to expedite the customer service experience for customers. Telecom companies also use virtual agents to drive customer service. By handling routine requests and providing customers with information, AI frees up human specialists for more demanding tasks. It’s worth noting, too, that the initial impact on workforce size has been minimal. Rather than reducing headcounts, it seems that companies are using their savings to invest in customer experience enhancements. This may change, however, as cognitive computing systems scale across sectors.
In any case, the most interesting thing about the data is not that companies in different industries implement the AI applications that best suit their value chains. What is interesting – especially for companies looking to make the most of the AI systems currently available – is that a suite of best practices has emerged across industries.
Success is About Skills
While CEOs used to worry about the availability of AI technology, concern has shifted to a more hands-on matter: skills. Many organizations are woefully unprepared to make the most of currently available cognitive computing systems.
While it may not be the next Industrial Revolution, it’s clear that AI really does open new frontiers for business. Yet knowing that and know-how are two entirely different matters. Today, it is know-how that drives AI success and failure, no matter the industry.
A suite of core practices are necessary to capture value at scale – and these practices apply across sectors. For example, businesses that outperform in the use of AI are significantly more likely to adopt it in business functions such as sales and marketing. What’s more, nearly 30 percent of high performers plan to increase AI investment by 50 percent or more in the next three years.
Underpinning these investments is an infrastructure that aligns strategy, IT, and data. Corporations intent on making the most of AI should develop data platforms, and it finds that almost half of the 12K+ organizations surveyed in the Global C-Suite study intend to, at a total value of $1.2 trillion.
Little Room for Prediction
Even though common themes cut across industries, rapid changes in AI adoption indicate that additional best practices may yet emerge.
Cognitive computing is no longer an emerging technology, but it is still far from enterprise scale. The 25 percent YoY increase in AI utilization in standard business practices indicates that this intermediary period may end quickly. Even so, a number of issues have yet to be resolved.
Few companies identify, monitor, and mitigate AI risk. Furthermore, regulatory frameworks often pose a barrier, rather than a support, to responsible innovation. Meanwhile, education systems the world over struggle to develop the human capital required to scale AI. That’s what makes this field so challenging to predict—and so profitable for those who predict correctly.
