Artificial Intelligence: Center of Continuous Improvement of User Experience

Original article was published by Pierre Pavlovic on Artificial Intelligence on Medium

Artificial Intelligence: center of continuous improvement of user experience

Photo by Peter Gombos on Unsplash

Artificial intelligence has changed the world of business and particularly the cosmetics market through a multitude of possible applications. Indeed, in his study for Juniper, Maynard, N. (2019) reminds that

“Global retailer spending on artificial intelligence services will reach $12 billion by 2023, compared to $3.6 billion in 2019. More than 325,000 retailers are adopting artificial intelligence during this period”.

Artificial intelligence capabilities are diverse and at the center of continuous improvement to enhance the user experience. In this article, we will focus on the significant limitations of artificial intelligence, but also the challenges that this technology can bring to the cosmetics industry through e-commerce.

Definitions and concepts of artificial intelligence

According to Saporta, G. (2019) “the mass of data created by companies, machines and individuals increases each year by 40%”. This exchange of very important data demonstrates the need to understand and implement new technological tools such as artificial intelligence to enhance the customer experience.

As defined by Soudoplatoff, S. (2018). “The best artificial intelligence engine is nothing without training. Investing in an artificial intelligence system means first and foremost having data sets that allow us to describe cases that have already been encountered”. Indeed the source of artificial intelligence is Data. If artificial intelligence is able to solve key points of the user experience, it is indeed thanks to the training of a large number of heterogeneous data (text, image, video, audio or digital).

According to Shapiro, S. (2000) “Artificial intelligence consists in understanding, designing and implementing ways to represent information in computers so that programs (agents) can use this information”. The goal according to him is to “ extract implicit information, converse with people in their natural language, decide on the future events, plan future activities, or solve problems in areas that normally require human expertise”.

Prototype of the use of AI for business by Charlie Isslander on Dribbble

Artificial intelligence seems to be today one of the major actors within the large companies of the cosmetics sector. It is a source for solving technological problems around the user experience with a multitude of possible applications, for simplifying processes or offering more efficient services to meet customer needs.

Artificial intelligence in E-Commerce

The use of artificial intelligence is especially important in the E-Commerce sector. Indeed, Linzbach, P., Inman, J. J., & Nikolova, H. (2019). define E-Commerce as “the measure of offering alternative and attractive value propositions with wide and deep assortments, practical purchasing processes and customized and tailor-made solutions”.

Thus this new technology, artificial intelligence, is completely in line with this value proposition. It is interesting to see what new technologies are under development.


Recently, chatbots have become an indispensable part of the E-Commerce ecosystem. Chatbots have an important deployment potential thanks to artificial intelligence technology as demonstrated by Ciechanowski, L., Przegalinska, A., Magnuski, M., & Gloor, P. (2019). “In fact, chatbots are a perfect example of the implementation of advanced consumer-oriented artificial intelligence that simulates human behaviour on the basis of formal models”. Thanks to this technology, e-merchants can offer a multitude of services directly to consumers on websites, social networks or in applications.

Initially used for advice, chatbots are now evolving towards monetization thanks to the understanding of user data. The first source of monetization is retargeting when a user abandons their shopping cart. Indeed, the user may be offered by the chatbot during a second visit the basket he had not confirmed.

Loyalty is the second aspect of monetization thanks to a more personalized loyalty, offers or discounts proposed by the chatbot. In the cosmetics sector, chatbots are most often used for advice and recommendations. Between human expertise and artificial intelligence, this is what the chatbot of the Helloava brand offers. This chatbot proposes a complete and detailed analysis of your skin type. Once the process is completed an appointment with a dermatologist is fixed and a pack of different items is proposed to you according to your choices.

Prototype for personalized skin diagnosis by Wout Helsmoortel for Bothrs on Dribbble

Even if this chatbot remains specialized in skin diagnosis, there are other types of chatbot notably rewarded for their efficiency. This is the case of Mitsuku, a conversational chatbot created by Steve Worswick, which is able to hold a natural and human conversation while learning from your data. Created by Facebook, Octave AI has also become a major player in the world of chatbots on basket dropping. Today, chatbots have become indispensable for consumer relations, and their objective is focused on personalization. Thus, some E-Commerce platforms go even further in the personalization of their platform through artificial intelligence.

Personalized recommendations

Speed and efficiency are nowadays the key words when a user arrives on an E-Commerce website. Indeed, the consumer needs to be met with new functionalities based on artificial intelligence. A multitude of projects are then offered to the platforms: virtual sales experience, personalized recommendations based on online or in-store purchases. The implementation of these solutions on E-Commerce platforms can then generate additional purchases and create a more personalized and proactive relationship with consumers.

Personalization is therefore a key point of the customer experience on an E-Commerce platform, as summarized by Eirinaki, M., & Vazirgiannis, M. (2003). “Web personalization is the process of adapting a website to the needs of specific users, taking advantage of the knowledge gained from analyzing the user’s browsing behavior (usage data) in correlation with other information collected in the context of the Web, namely structure, content and user profile data”.

Prototype of personalized recommendations by Gleb Kuznetsov on Dribbble

The integration of artificial intelligence in the cosmetics sector is a vector for new innovative functionalities. One of these examples is the use of virtual fitting of cosmetic products. This new type of use of artificial intelligence allows consumers to virtually try on items before purchase. In 2018, L’Oréal purchased Modiface, a company specializing in augmented reality and artificial intelligence. Modiface is a complement to Vichy’s SkinConsultAI and becomes the first skin diagnostic application offered by L’Oréal. This technology provides women with a skin diagnosis by following certain steps. Consumers are invited to take a photo on the site. The technology will detect different signs of skin damage and the consumer receives a personalized matrix with information about the aging and strengths of their skin, followed by a recommendation of suitable products. Another application of this technology, Modiface now allows Lancôme consumers to test a virtual make-up experience based on augmented reality.

The progress of artificial intelligence in the cosmetics sector has been growing strongly in recent years. This technology can redefine all online strategies by offering new services to the consumer. A technology that is sometimes frightening and a subject of discord, is artificial intelligence a new challenge for the cosmetics sector?

A new challenge: Adapting and adopting artificial intelligence

In his book, “Artificial Intelligence Doesn’t Exist” published in 2019, Siri co-creator Luc Julia points out the lack of information about artificial intelligence “My biggest fear is that people will believe everything they hear about AI, simply because they don’t have enough information on the subject. The whole point of a good education is to acquire enough knowledge to develop a critical mind”.

Indeed, we are entitled today to ask ourselves the question of this new challenge on the adoption and adaptation of artificial intelligence among users.

The advantages

According to Amellal, S. (2020). “The change of organizations towards artificial intelligence is inevitable”. The challenge is budgetary, because companies that are not sensitive to this technology in the future will no longer be competitive. The advantages of artificial intelligence in the cosmetics sector are multiple. According to the study by Chui, M. (2017) for the McKinsey Institute 3 axes of benefits emerge.

  • Prediction using deep learning to predict purchases or trends in e-commerce or in-store. Product stocks can be reduced by up to 20%.
  • Gain in productivity, artificial intelligence allows us to gain in productivity by using autonomous vehicles and space planning, stocks and therefore surfaces can be reduced by 30%.
  • Using personalization and dynamic prices online sales can increase by 30%. Moreover, this personalization can also be expressed through the use of artificial intelligence to modulate and manage the user interface and thus increase sales by 4 to 6%.

Even if artificial intelligence can represent an important investment cost, it has a number of advantages that can improve conversions. Despite this, this technology has some notable limitations.

The limits

Despite the considerable advantages offered by artificial intelligence, certain points remain unresolved, including the maturity of this new technology. According to LeCun, Y., Bengio, Y., & Hinton, G. (2015). “Eventually, major advances in artificial intelligence will be achieved through systems that combine learning representation and complex reasoning”. Indeed, before using artificial intelligence, it is necessary to be able to have a certain amount of reliable and usable data to be able to train the algorithms and in particular to avoid overlearning.

Another technological concern today is the capacity of computers and transistors to meet expectations by following Moore’s law, which according to Loeve, S. (2015). “is mostly cited at times when (as today) one evokes its probable end with the Moore’s Wall’’.

However, even if the technological limits are present there are also environmental limits, with the use of massive energy resources for the training of algorithms. But there are also ethical limits, notably on “The possibility of creating thinking machines that raise a host of ethical questions” described by Bostrom, N., & Yudkowsky, E. (2014). But also on the use of people whose goal is to label or verify the decision of a censoring algorithm.

Thus, even if artificial intelligence is now increasingly adopted by companies, it is sometimes not adapted to certain usages. Indeed, Fetzer, J. H. (2012). notes “The problems we face with the scope and limitations of AI are complex”. Artificial intelligence leads us to changes in our organizations but also in the way we work with a significant cost for companies.


Today, companies have the opportunity to implement many things related to artificial intelligence to improve the user experience. Indeed, we have seen that users are looking for new experiences that delight them or simply save them time. Thanks to artificial intelligence, companies have the opportunity to enchant the customer. As we have seen, artificial intelligence allows them to differentiate themselves, but is it the only technology that can improve the user experience?

Sources :

Saporta, G. (2019, May). Science des données, données massives: défis et nouveaux métiers. In CISEM 2019.

Soudoplatoff, S. (2018). L’intelligence artificielle : l’expertise partout accessible à tous.

Shapiro, S. C. (2000). Knowledge Representation and Reasoning Logics for Artificial Intelligence.

Linzbach, P., Inman, J. J., & Nikolova, H. (2019). E-Commerce in a Physical Store: Which Retailing Technologies Add Real Value? NIM Marketing Intelligence Review, 11(1), 42–47.

Ciechanowski, L., Przegalinska, A., Magnuski, M., & Gloor, P. (2019). In the Shades of the Uncanny Valley: An Experimental Study of Human-Chatbot Interaction.

Eirinaki, M., & Vazirgiannis, M. (2003). Web mining for web personalization. ACM Transactions on Internet Technology (TOIT), 3(1), 1–27.

Julia, L., & Khayat, O. (2019). L’intelligence artificielle n’existe pas. edi8.

Chui, M. (2017). Artificial intelligence the next digital frontier?. McKinsey and Company Global Institute, 47, 3–6.

Lecun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. In Nature (Vol. 521, Issue 7553, pp. 436–444). Nature Publishing Group.

Loeve, S. (2015). La Loi de Moore. Enquête critique sur l’économie d’une promesse. Sciences et technologies émergentes: Pourquoi tant de promesses, 91–113.

Bostrom, N., & Yudkowsky, E. (2014). The ethics of artificial intelligence. The Cambridge handbook of artificial intelligence, 1, 316–334.

Fetzer, J. H. (2012). Artificial intelligence: Its scope and limits (Vol. 4). Springer Science & Business Media.

Nick Maynard. (2019). How AI can revive retail.