28 october 2020
AI used in business: How a machine helps you “make” money

AI used in business: How a machine helps you “make” money

Only ten percent of companies have been able to increase revenue by adopting artificial intelligence. Researchers have figured out what lies at the heart of their success.

Among companies using artificial intelligence in workflows, only a tenth was able to reap significant financial benefits from this, according to a report from analysts at the Massachusetts Institute of Technology and the Boston Consulting Group.

Human-machine interaction

Analysts at BCG, one of the “Big Three Management Consulting” and the research journal of the world’s best technical university MIT Sloan Management Review, published a study this week on the results of AI in company workflows.

In recent years, artificial intelligence has been introduced in a variety of areas, including those far from IT technologies. However, only ten percent of them managed to use the algorithms with financial benefits for themselves.

The researchers surveyed more than 3,000 managers from 29 industries and 112 countries. Most of them (57 percent) reported that their companies are testing or already implementing artificial intelligence in their workflows. 59 percent said their companies have developed an AI strategy.

At the same time, more than 70 percent of the managers surveyed reported that their company has an understanding of how AI can help to generate profit from the business. The report notes that three years ago, 57 percent of respondents agreed with this question.

However, only one in ten executives who took part in the study said that the introduction of AI has already brought them significant financial benefits.

The study authors conclude that the key to success is a combination of human and AI interaction. The effectiveness of artificial intelligence in such a situation can increase six-fold.

Analysts have identified five ways in which algorithms and humans interact:

• AI makes decisions and implements them

• AI makes decisions, and humans fulfill

• AI makes recommendations, and humans make decisions

• AI generates ideas and analytical conclusions, and people use them in the decision-making process

• people generate ideas and conclusions, and AI evaluates them

Among companies that use only one of these methods, only five percent have financial success. If there are two ways – six percent.

When using three or four options, it is 15 percent, but when all five interactions are combined, it is 32 percent.

“The single most important factor that determines value creation using AI is not algorithms or technology, but a person. A select group of successful companies are better than others at creating integrated AI systems – a person in which AI learns from a person, and humans learn from AI, “says Max Hauser, Managing Director and Partner at BCG.

Artificial intelligence is now called neural networks that use the so-called deep learning method. Such a network in some way reproduces the organization of neurons in the brain, and in fact is a set of simple calculators connected with each other and able to receive input data, transmit signals to each other and form a response. … The more complex the architecture of a neural network, the more complex problems it can learn to solve.

One of the most promising uses for machine learning is medicine, namely the search for new medicine.

So, at the beginning of this year it became known that scientists at the Massachusetts Institute of Technology for the first time in half a century have found several new antibiotics. A neural network helped them make a discovery, which was a unique achievement in the history of medicine.

And in October 2020, MIT presented a model that could generate new drugs for tuberculosis. Most of the variants of the algorithm turned out to be effective.

This is due to the introduction of a new feature in machine learning algorithms that improves the ability to predict.

Using a new approach that allows computer models to account for uncertainty in the data, the institute’s team has identified several promising compounds that target a transport protein required by M. tuberculosis bacteria. If it is absent or inactive, then bacteria can no longer multiply.

Another interesting application of neural networks was reported at the end of September. Then the New York authorities resumed the work of a pilot project on the use of artificial intelligence in the judicial system.

Authorities hope AI will help them reduce the burden on local prisons and avoid the bias that characterizes conservative American judges.

Subscribe to our newsletter

By clicking the button, I accept the terms of the Offer for the use of the site and agree with privacy policy