This article can be seen as a logical continuation of my previous article "Do you need to use artificial intelligence in your business?".

Now, artificial intelligence (AI) is almost everywhere we look:

1. Chatbots are able to communicate with customers at a deep level and can even respond appropriately to the emotional color of the interlocutor's messages.

2. Walmart has long been introducing robots designed to scan warehouses for scarce stocks, missing products and inaccurate prices.

3. Another well-known example is the Amazon Go store, where the purchases are organized exactly as it should be in the ideal world of self-service: you browse the store, choose the right products, put them in a bag, a backpack or just put them in your pockets. Then leave the store by special turnstiles, after which the purchase amount is automatically debited from your Amazon account.

Artificial intelligence is practical and speeds up decision-making on a much larger scale than expected in retail.

However, even with the increasing level of comfort in different areas of life and improved customer service in business, not everyone is aware of the latest changes. Some fear that AI can replace human intelligence. Recently, it has often been found in the newspapers that Elon Musk predicts that AI will one day be a danger to human civilization. Nobody can predict the distant future, but for the moment, the decision-making process exists in a profitable symbiosis relationship between a person and an AI-based machine, but controlled by people.

At the moment, artificial intelligence is very effective at extracting data from multiple sources, enriching it, classifying and generating information quickly. The retail industry has used the AI to provide online advice, help customers find the right products and even anticipate the needs of their customers. For example, KFC is testing software to scan visitors' faces, assess their mood, age, etc. Based on all this information, the AI tries to offer the visitor dishes he might like. In other words, the goal of the project is to provide an individual approach to each visitor. If the customer is a frequent guest of KFC, the IA analyzes the visitor's previous orders and offers a choice of dishes according to their preferences.

Although artificial intelligence has many advantages, it also has drawbacks: AI does not work when objectives are not clearly defined or when multiple conflicting results need to be evaluated. It also does not take into account non-quantifiable or intangible data, such as reputation or branding. In these cases, the judgments of human intelligence will be better than AI. That's why human-machine interaction is the most effective strategy. The combination of human and artificial intelligence allows you to make more informed and reasonable decisions.

Decision making is better when own specific data is used

AI is as strong as its data. That's why it's important to make sure the distortion of the system has been removed from the data. This will help separate the necessary signals from the noise.

Take the example of a B2B distributor who, after a major acquisition, offered a wide range of products. The company needed to rationalize its range. Unfortunately, they could not accurately point out to their customers replacement products for those who were no longer available. This led to the fact that many customers could not find the products that suited them. No alternatives were offered for 80% of out-of-stock products, which resulted in a low customer retention index (NPS), net promoter score, and a 10% drop in sales.

To find a solution, experts worked with analysts to help them understand each product category. After receiving, visualizing and mixing the data, based on comments from the category manager, various algorithms correspondence have been created. At each iteration, the algorithm has been further improved to ensure better product consistency. This "man + machine" approach helped increase the coverage of alternative products from 19 to 78%, improve the quality of customer service and increase sales by 5%.

Today, the collaboration of experts, artificial intelligence and data can accelerate the development of the sector, but can also be destroyed by lack of skills.

Artificial intelligence no longer means "artificial intelligence," but "augmented intelligence," which is a reasonable combination of man and machine.