Daily Management Review

Study: AI is not as profitable as you might think


A global survey of the structures of the University of MIT and the BCG group confirmed that artificial intelligence (AI) technologies entered the boundary zone of mass application in business in 2019. The vast majority of companies are confident that AI will play a big role in business; almost half of them see the risk that AI will become a factor in the success of competitors. However, not more than 30% of them talk about existing influence of AI on their business, and less than 40% of those who invested in it made a profit. The authors believe that the successful use of AI without adequate integration into corporate strategy is impossible.

The Media and Research Unit of the Sloan University of Massachusetts University of Technology's MIT Business School (MIT Sloan Management Review) and the AI Development Division (BCG Gamma) of the AI Boston Consulting Group published the AI Success Report. The global (97 countries) survey covered more than 2500 representatives of large companies in 27 industries (including insurance, IT, banks, industry, healthcare, pharmaceuticals, retail and mining). Analysts at Sloan and BCG apparently managed to catch an unusual moment in the study of business perception of a potentially promising technology that had been developing in the academic environment earlier for several decades. Existing AI implementations are already perceived by companies as reality, but most of the companies investing in AI demonstrate its effectiveness unable.

There is consensus in assessing the future importance of AI: nine out of ten respondents (top management, including non-core) are confident that it will significantly affect the business environment in the future. Meanwhile, less than 40% of these investments have brought commercial benefits for the majority (90%) of those who have already invested in AI over the past three years. The paper includes data from companies that “use” external AI technologies and develops them internally (there are more of them), and also contains estimates of companies that use AI to reduce costs (an ultrapopular direction that mainly saves on labor costs and equipment), sales growth and creation of new products. The work clearly demonstrates a certain tension around AI as a technology.

Here are the brief findings of Sloan and BCG regarding the most efficient and commercially successful use of AI in companies. AI strategies are less successful when they are fully integrated into the general digital strategy and in the management strategy, including in business transformation projects. AI for sales growth is more profitable than for lowering costs. “Inside production of AI” is better than its “consumption”. Finally, AI should be coordinated with the company's data-strategy and innovative personnel policy. “In cases where AI initiatives are managed by the IT director, the chances of seeing real benefits from the introduction of new technologies are 50% lower,” the review suggests. Note that the low efficiency of AI in non-modernized, including managerial, business was theoretically predicted in academic circles over the past 15 years.

source: sloanreview.mit.edu