IT trends in retail

Author: Boris Agatov, independent expert on retail innovation.

Hyperautomation

In a very short period of time, there has been an evolution in the understanding and use of AI in retail.

Before the advent of public generative AI, it was expensive, difficult to implement, and not very reliable in its results. If a company implemented AI tools, then it tried to entrust it with the most important tasks (inventory optimization, sales forecasts, etc.), because budgets were so high that they did not allow it to be used in routine, non-key low-level tasks. But with the advent of public AI, the picture begins to change rapidly, and the understanding that AI can be entrusted with a wide range of low-level tasks, because... it has become available and is no more difficult to use than Excel.

It still has problems with accuracy (hallucinations) and reliability, but this is not an obstacle to its implementation in low-level processes, because they are easy to double-check for a human. For example, you decided to introduce an assistant with AI that will help employees correspond with contractors. Such implementation will make it possible to increase productivity by two or more times, but if suddenly the AI ​​writes an erroneous letter, a person will quickly notice this and make adjustments. It’s a different matter when complex analytical calculations are made using AI, for example, sales plans for the year, and there is no way to check them as easily as a letter.

A huge number of lower-level business processes are subject to automation according to this principle: mass automation is quite capable of leading us to the phenomenon of hyperautomation, which will increase our efficiency significantly. According to the law of dialectics, we automate everything, and quantity turns into quality.

AI assistants (co-pilot), AI agents changing IT infrastructures in retail

There are indications that IT infrastructures may undergo significant changes with the advent of public AI.

What awaits us in this direction?

The entire IT infrastructure will become sharply smarter. If we imagine the standard path of information passing through the IT infrastructure: information input, processing, information output and information transfer. There is a lot that is familiar to us on this path, but there are also many moments that we would like to tweak a little, but this is almost impossible.

For example, we would like to enter information in any form (video, sound, text) and of any quality, and output it not in standard report forms, but in an arbitrary form that will allow us to solve business problems more efficiently, for example, we need not table, and a video for employees, now such an approach is impossible to imagine.

AI assistants will stand in front of each block and help us switch to the natural language of communication with the IT infrastructure to solve a business problem. Such an AI assistant will be able to help collect, transform and clean up information for input. Another AI assistant will help you display information in any desired form. Such AI assistants will make IT systems responsive to any business tasks.

The next stage in the development of AI assistants will be the emergence of AI agents. These will already be entities that will be endowed with the right to make decisions and interact with each other. Now this is the latest trend in automation and achieving autonomy of entire business blocks, from procurement to HR tasks. With the advent of AI agents, business processes will change radically, because... a lot of them will be given to AI agents, and there will be no need to involve a person.

But the first stage will be numerous AI assistants (co-pilots), which will work together with a person, under his control for his convenience, but then such systems will be replaced by autonomous AI agents.

Hyper-situationalism and hyper-personalization of content

Generative AI has given us the ability to create unlimited amounts of video, audio, image and text content at very little cost. Nowadays, making a personal video for each customer conveying an individual offer is not a prohibitively difficult and expensive task. Moscow clinics send out videos where the doctor addresses them by name, mentions their age, chronic disease, and invites them to see a specific doctor. At all stages, AI is used, which will voice the text and make the correct articulation of the lips so that it looks incredibly natural.

Nowadays, creating a digital avatar of any character costs almost nothing, and can be done in literally 5 minutes. I am already actively recording videos with my digital avatar, which voices the content in my voice and with my image, which the AI ​​helps me prepare. Creating such a video with live subtitles costs about 200 rubles, i.e. almost nothing, and allows you to stream video content creation.

It seems that there is no need to digitize anyone, because... The creation of digital celebrities is gaining popularity, who easily and naturally lead communities, communicate with everyone individually and do not have any weaknesses that could, unlike a living person, steal his reputation.

This way, we can achieve hyper-personalization in the creation of virtually any customized content for any number of customers.

AI tools allow you to instantly change images, videos and sounds - this will help to play out changes in any external circumstances, for any volume of marketing visual materials. For example, you are selling an autumn collection, and all marketing images (videos, banners, etc.) are in autumn exteriors. And suddenly the snow fell earlier than usual, and winter set in. Using neural networks, turning autumn into winter is no longer a difficult task. Thus, we come to hyper-situationalism in marketing, when any situation is supported by marketing materials appropriate to this situation and leads to an increase in sales.

Retail media network

Retail Media Network (RMN) is an advertising platform built by retailers that allows brands to advertise directly on retailer channels.

It combines the different advertising channels that you have available:

• Store website
• Mobile app
• Email newsletters
• Social media
• Screens in the store
• Checks

These are like billboards, but in an online store, where the advertisement is shown to customers who are already interested in purchasing.

Imagine that you are a store owner. There are many visitors to your store who have come to shop. Now imagine that you offer other companies the opportunity to place their advertising directly in your store, where it will be seen by many targeted customers. This is the essence of Retail Media Network.

Example:

A chocolate bar manufacturer wants to increase sales.

You, as a retailer, can offer the manufacturer to place advertising for the bars on screens in your store, e-mail newsletters, etc.

The advertiser pays you for impressions or clicks on the ad.

The most striking example of this kind of platform is marketplaces, which earn up to 10% of turnover from posting such content (OZON’s income from advertising was 24 billion rubles).

What else is important to know about Retail Media Network (RMN):

This tool is now approaching the reach of television.

The growth of this channel is 20% per year.

Retail Media Network is quite difficult to manage, because... Each retailer has its own advertising space formats; it is difficult to keep track of expenses, because... you need to conclude many contracts, problems with impression metrics. But, despite such difficulties, this is one of the most promising channels in the world, and especially in Russia, against the backdrop of the loss of many promotion channels on social networks. The popularity of this channel lies in the fact that there are buyers who have already come to buy something, and the likelihood that advertising will be more attractive to them is higher than if a person was watching some kind of entertainment content and there was an advertisement for the product.

Among the latest interesting developments, one cannot fail to note the emergence of humanoid robots.

So Figure signed its first commercial contract and is sending its general-purpose humanoid robots to work at a BMW plant in South Carolina.

In accordance with the recently signed commercial agreement with BMW Manufacturing Co. LLC, Figure began identifying initial use cases at BMW's Spartanburg plant and began training bots for phased on-site deployment. Figure's deal with BMW is purely commercial, which could make it one of the first, if not the first, deal of its kind.

Also interesting is the experience of autonomous humanoids working side by side with people - Amazon, for example, announced in October 2023 that it will begin testing Agility Robotics' humanoid Digit as a "mobile manipulator" carrying containers in warehouse situations where there is not enough space for conveyor belts. belts—and in older facilities that were not yet specifically designed with flat floors suitable for wheeled robots.

Amazon now has more than 750,000 robots that work alongside employees to perform repetitive tasks and free up employees to be more effective at serving our customers.

Finally

Our society, and with it retail, faces an incredible challenge, which is associated with the accumulation of a critical mass of new technological approaches, and, first of all, publicly available intelligence. Testing and developing methods for implementing new technologies will allow companies to be sustainable and customers to have a new and improved shopping experience. A lot will change very quickly under the influence of new technologies, so it is necessary to pay attention to retraining staff and changing the corporate culture, which will allow many to remain in demand during this technological transition.

Russian retail, from the point of view of customer experience and technological equipment at the global level, looks quite confident and occupies leading positions in many positions, so it is so important not to miss the trends that I described above, so as not to lose our leadership.