Artificial Intelligence and Machine Learning: Should a brand that aims to be close to its customers and communicate with them in their language as its main goal use algorithms for this task?
We are captivated by stories of exceptional customer service, where the brand cares about a personal relationship and works hard for every positive feedback. On the other hand, all business areas are currently moving towards automation. So how to ensure high quality of the customer’s personal experience with the brand when the human-human relationship transforms into a human-algorithm relationship? This question was answered by the consumers themselves, talking about their attitudes, feelings, and beliefs about artificial intelligence.
Getting Start With Artificial Intelligence And Machine Learning: Reducing Frustration
In the past six months, OMD, in collaboration with the University of London’s Goldsmiths College research team, conducted a project to identify the opportunities, benefits, and risks of implementing artificial intelligence in sales and customer service. The aim of the study, which covered 12 European markets, was to find out about current consumer preferences, similarities, and differences in the perception of artificial intelligence, consumer behavior, and how artificial intelligence is used. Analyzing the level of tolerance, understanding, and consumer confidence in artificial intelligence has determined how brands can successfully implement AI-based services.
Twenty-two percent of adult Europeans already use AI-based applications or devices. In Poland, this percentage is slightly lower (16%), but the potential to disseminate this type of solution is highly rated – 58% of Poles would like to have a device or application using artificial intelligence – this result is well above the European average.
Discover even more and more often. You guarantee yourself access to a solid dose of Harvard knowledge by purchasing a subscription. Make sure you won’t miss anything.
The most attractive functionalities for consumers, which would be supported by artificial intelligence, meet basic needs, such as payment for purchases, product testing without the participation of the seller, tracking the delivery status, or offering new products and services. The need to “package” these functionalities in an artificial intelligence algorithm may result from the fact that the areas indicated by customers are also the source of their daily frustration.
The five most common reasons you can get frustrated when shopping are :
- problems with the shipment,
- difficulties with the return or exchange of goods,
- waiting in line,
- purchasing pressure exerted by the seller,
- no information on the quality of products purchased online.
The algorithms’ main task should be to ensure maximum smoothness of the process and immediate provision of information to the customer – when it is needed.
What About Building A Relationship?
In the conditions of unlimited choice, consumers appreciate the benefits of various types of advice designed to help them make the right decision. Research shows that when shopping in the food, cosmetics, books, and furniture categories, customers would be willing to use help in order to make a more informed decision. In other categories (e.g., travel, movies, and TV), they would like to surprise by new possibilities and ideas. On the other hand, buyers are open to personalized suggestions when it comes to beauty, health, and fitness.
The above-mentioned situations provide an important opportunity to build relationships with customers. However, this does not mean that in every case, the client’s advisor must be a human being. Artificial intelligence is increasingly becoming a source of support in the decision-making process. Base on the analysis of habits and preferences. And then referring them to a group with similar purchasing behavior, the algorithms are able to choose the right product for the customer with a high probability of success.
Relationship Between Artificial Intelligence And Machine Learning
Concern about the way the obtaining data will use for the two main reasons for the lack of trust in artificial intelligence and machine learning.
Artificial intelligence makes the best decisions base on user data, but not all customers are willing to share their data to the same extent. An OMD survey shows that 42% of consumers would consent to share their data with an AI-powered platform. Concern about how the acquire data will use for the two main reasons for mistrust of AI. As many as 40% of consumers are afraid that their data will use in an unfavorable way.
While 47% indicate the lack of control over the entire experience as the main barrier to trust. However, it is worth noting that despite little enthusiasm for sharing data on the AI platform. Some consumers may not realize how much information they already provide about themselves. E.g., on social networks, on music services such as Spotify. Or in online rental companies type of Netflix.
If A Bot, It Is Informative & Non-Invasive
Is a virtual advisor in the form of a chatbot able to meet the expectations of users as much as a human being. At the same time creating relationships by assisting in the world of purchases? It depends on what the bot has to offer to the customer.
If it is an automated version of a seller strongly persuading to buy. Then its chances of success are slim. But if it can initiate engaging interactions by providing valuable information for the customer. Or suggesting the best solutions during purchasing dilemmas. It may prove to be a reliable advisor or even guardian.
An example is a Recipehelper chatbot design by the OMD project. And the University of London’s Goldsmiths College for the British supermarket chain Waitrose. The recipe helper generates many different recipes base on the recognition of the images of the products in the user’s refrigerator.
The customer takes a photo of the contents of the refrigerator. The algorithm uses Watson Image Recognition software, identifies available ingredients. And suggests meals that can prepare or suggest a culinary inspiration. When testing the chatbot with the respondents, they liked most that the bot. Will ask how much time they have to prepare a dish!
If A Bot!
Another aftermarket solution will design for the John Lewis department store chain. Chatbot Sleepbot will create for customers who will buy a bed with home delivery. Remind about the delivery date, offer help in the form of bed assembly instructions, and will suggest relating products. During the chatbot tests, the suggestions of relating products seem too intrusive to customers. While the respondents appreciate the functionality of the delivery reminder.
In addition to chatbots with sales and service applications, attempts will do to use their functionality to handle complaints. In this aspect, the key is an accurate assessment of the type of problem. And appropriate redirection for machine operation or human intervention.
Consumers’ moods show that the potential of using artificial intelligence and machine learning in the purchasing process is high. And solutions such as chatbots can prove successful at individual stages of the purchase and trigger spontaneous interactions with the brand.