Artificial intelligence has become one of the biggest technological developments in business in recent years, but the field is still largely shrouded in uncertainty. While expectations run sky-high, what are businesses actually doing now? A new report by BCG and MIT Sloan Management Reviewaims to demystify AI in business and take stock of current industry adoption. The report is based on a global survey of more than 3,000 executives and in-depth interviews with more than 30 technology experts and executives. Its goal is to present a realistic baseline that allows companies to compare their AI efforts and ambitions and to provide guidance for things to come.
The high expectations for AI cross geographies, industries, and companies, regardless of size. While most executives have not yet seen substantial effects from AI on their offerings and processes, they clearly expect to in the next five years. (See Exhibit 1.) Most organizations foresee sizable effects on IT, operations and manufacturing, supply chain management, and customer-facing activities. Business-process-outsourcing companies, for example, expect many of the jobs that moved to low-labor-cost countries in recent years to be automated. However, they also expect AI to lead to new activities and sources of employment. Executives at industrial companies expect the largest effect in operations and manufacturing. In connection with its new A350 program, for example, Airbus is using AI to speed and improve production. The company has combined data from past production programs, continuing input from the A350 program, fuzzy matching, and a self-learning algorithm to identify patterns in production problems. In some areas, the system matches about 70% of seemingly unrelated production disruptions to solutions used previously—in near-real time.
AMBITION AND EXECUTION
Expectations aside, the gap between ambition and execution is large at most companies. Three-quarters of executives believe that AI will enable their companies to move into new businesses. And almost 85% believe AI will allow their companies to obtain or sustain a competitive advantage. But while more than 60% of respondents said that a strategy for AI is urgent for their organizations, only half of those said that their organizations have a strategy in place. The largest companies (those with more than 100,000 employees) are the most likely to have an AI strategy, but only half have one.
There are also large gaps between leaders and laggards. Among leaders, three-quarters have identified business cases for AI. About 80% say that senior leaders are onboard. The biggest hurdles at these companies are hiring and developing talent and establishing priorities for AI investments; they are also starting to worry about security issues related to AI. Laggards, on the other hand, have not identified business cases. More than 50% report that their senior leaders are generally not involved in AI, and most have yet to encounter the difficulties of sourcing AI talent.
The differences in adoption can be striking, particularly within the same industry. Ping An Insurance Company of China, one of that country’s largest insurers, employs 110 data scientists and has launched about 30 CEO-sponsored AI initiatives that support, in part, its vision “that technology will be the key driver to deliver top-line growth for the company in the years to come,” says the company’s chief innovation officer, Jonathan Larsen. Elsewhere in the insurance industry, AI initiatives are at the other end of the spectrum, limited to such efforts as “experimenting with chatbots,” as a senior executive at a large Western insurer described his company’s AI program.
DATA, TRAINING, AND ALGORITHMS
One of the most telling differences between leaders and laggards is their understanding of the importance of data, training, and algorithms. AI algorithms are not natively “intelligent.” AI starts with “naked” algorithms that become intelligent only upon being trained on large amounts of data and, for most business applications, large amounts of company-specific data. Successful training requires fully understanding this process and the role of data, which is far more significant than it is in big data and advanced analytics applications. Success also depends on having well-developed systems that can pull together relevant training and continue to integrate findings from later data collected over time. Data collection and preparation are often the most time-consuming activities in developing AI applications.
Thus, even if the organization owns the data it needs, fragmentation across multiple systems can hinder the process of training AI algorithms. Agus Sudjianto, executive vice president of corporate model risk at Wells Fargo, put it this way: “A big component of what we do is dealing with unstructured data, such as text mining, and analyzing enormous quantities of transaction data, looking at patterns. We work on continuously improving our customer experience as well as decision making in terms of customer prospecting, credit approval, and financial crime detection. In all these fields, there are significant opportunities to apply AI, but in a very large organization, data is often fragmented. This is the core issue of the large corporation—dealing with data strategically.” Less than half of our survey respondents said that their organization understands the data needs of algorithms or the processes required to train algorithms.