AI to help brands with content creation for enhancing customer experience.

Introduction

In the article Data Science touches Customer Emotions, we discussed how emotions and sentiments are important for customer experience and how they are related to brand trust and brand love. Further, we discussed how deep learning capabilities as modern AI can be used to quantitatively analyse emotions and sentiments. This article reviews the importance of customers' emotional engagement with the brand, how emotional engagement is driven by emotional content, and how AI can help to create emotional content for enhancing customer experience.

Background and Motivation

The quantitative analysis of emotions is still an emerging field in science and technology. We are developing an open-source deep-neural-network based emotion-analysis framework which can be cloned from GitHub:

https://github.com/samarawickrama/AI-Image-Classification,

which was adopted from GoogleNet Inception-V3 [1] convolution neural network (CNN). For demonstration purposes only, the framework was customized to quantitative analysis of facial emotions, which can be tested from

https://52.63.103.27:5000

The neural network has been re-trained with the AffectNet-emotion database and some other online images. Please refer to the paper [2] for more information about the AffectNet-based emotions.

In this review, we will investigate beyond AI's emotion classification capability, and analyse how such a capability can be used to enhance the customer experience by creating relevant and emotional contents for brands. The customers' emotions on brand contents are complex, amorphous and dynamic which has many dependencies. For example, it differs from brand to brand and from time to time. The same content published in different brands creates different emotions. Further, the same content published in the same brand in different times creates different emotions. Therefore, a brand needs a dynamic content-personalization framework to create the best content for their customers.

How are brand advocacy, emotional engagement and contents mutually related?

Brand advocacy requires strong emotional engagement

In customer-brand interaction, brand love is which gives passionate affection that consumers have towards a brand [3], while brand trust is a feeling of security held by the consumer in their interaction with the brand [4]. Therefore, consumer emotions and customer experience have a strong connection. To improve customer experience, the brand should enhance the customers' emotional engagement with the brand.

Many kinds of research have been conducted to analyse social media engagement and brand experience. In the paper [5], how Facebook reactions can be mapped to customer journey was modelled. As shown in the following figure [5], different positive Facebook reactions (i.e., the content engagement) reflect customers' level of cognitive and emotional engagement with the content, which can be incorporated to improve brand advocacy.

As an example, it shows that Facebook sharing has more advocacy than liking on to a post. Based on the sharing and liking, it is possible to estimate the customers' brand experience.

The emotional engagement is driven by emotional content

In consumer-content engagement, emotional contents get more attention and drive the engagements [6]. Therefore, the content-design strategy should have a focus on touching the emotions and feelings of the customers. In this perspective, some research [6] emphasizes that there should be a neurophysiological method to capture the complex emotion construct in the contents to improve the content-engagement process. The storytelling in brand marketing is also another concept for enhancing the customers' emotional engagement with the brand contents [7]. Following is a video explaining how emotional and personalized contents drive the Coca-Cola brand experience.

Why does a brand need a personalized strategy for contents?

Kashdan et al. analyzed the diversity of human curiosity in their research [8]. Curiosity is one of the cognitive functions which can be used to improve the quality of brand and social-media content. As per the paper, related to a subject of interest, there can be four types of curiosities such that:

1. Fascinate: The curiosity which drives human social, enthusiastic, assertive, and aspiring others. This thrives life as an adventure and unpredictable. Further, it drives passion which translates into wide-ranging expertise. This curiosity depends on the individual's education and affluence on the engaging topic,
2. Problem Solve: The curiosity which drives human hard-work. This curiosity emerged with the independence and feeling of love to solve the problem. This curiosity group does not tend to ask a lot of questions and shows less interest in luxurious activities such as accessing social media, magazines, and less interested in interactions,
3. Empathize: The curiosity which drives humans to know what makes them inspired. Despite driving social perceptive, this curiosity creates attention to what is going on around instead of driving participation,
4. Avoid: This is the least curious behaviour among confident, educated and affluent individuals. It makes shyness on things which cannot be understood. Further, It creates stress more often than any other curiosity, avoiding emotional engagement with the content. Consequently, this creates fewer passionate interests on the topics.

Moreover, the association of these curiosities to the audience is important as each curiosity group has similar probabilities to exist in the population (see the following table) while being loosely correlated to each other (from Table 2 in [8]).

Therefore, for delivering a wider consumer experience, it is important to address these diversities by personalising the contents based on emotions and cognitive aspects of the customers.

How can AI help to uncover the emotional attributes of the contents?

Intuitively, deep learning is inspired by the artificial neural network which is the same concept the brain works. In general, this is what is referred to as modern artificial intelligence (AI). Deep Learning is the discipline which thrived AI long way recently enabling capabilities like self-driving cars, humanoid robotics and big-data analytics. While there are many disciplines and applications of deep learning which can be used for content creation, here the image processing and natural language processing (NLP) capabilities are discussed, as they are highly relevant for brand and social-media content creation/evaluation.

Image Processing

A picture is worth a thousand words, which expresses how important images are when engaging with customers. Facebook, Instagram, brand websites and many other channels usually grab customer attention by images. Following block diagram [5] shows how image attributes drive cognitive processing and drive consumer responses.

Modern deep-learning-based AI can be used to quantitatively measure the emotional attributes of the images. These measures on past-image contents together with their respective customer responses can be used to develop a customer-content-engagement model. Following are some of the widely used cloud deep-learning-based image-processing APIs which can be used to analyze emotional attributes of the images.

Following is how image-processing-based AI can be used in customer engagement with attention to customer emotions:

Natural Language Processing (NLP)

Similar to images, the richness of the text contents is also very important when emotionally engaging with the customers in channels like Emails, Direct mails, Website, Twitter, SMS, Facebook, etc. The deep-learning-based natural language processing (NLP) can be used to analyze the sentiments, semantics and emotion of the text contents to enhance the quality of the communications. Following are some of the widely used cloud deep-learning natural language processing (NLP) APIs which can be used to analyze sentiments and emotional attributes of the text contents.

Following is an example of how natural language processing (NLP) can be used in customer engagement with attention to customer emotions:

Conclusions

The customer emotions on brand contents are complex, amorphous and dynamic which have many dependencies. However, this mapping is an important aspect of creating customer-driven brand and social-media contents. The marketing researchers explain, a successful customer experience needs a personalized content strategy which accounts cognitive diversities. The modern AI which is inspired by deep neural network based machine learning can support this content personalization by mapping emotions to brand contents. The deep-learning-based image processing and natural language processing (NLP) are 2 of the disciplines evolved recently which can be adopted to enhance the brand & social-media content to improve the brand experience with attention to customer emotions.

References

[1] Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J. and Wojna, Z. (2016). Rethinking the Inception Architecture for Computer Vision. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Available at: https://arxiv.org/abs/1512.00567.

[2] Mollahosseini, A., Hasani, B. and Mahoor, M.H. (2019). AffectNet: A Database for Facial Expression, Valence, and Arousal Computing in the Wild. IEEE Transactions on Affective Computing, [online] 10(1), pp.18–31. Available at: https://arxiv.org/abs/1708.03985

[3] Hegner, S.M., Fenko, A. and Teravest, A. (2017). Using the theory of planned behaviour to understand brand love. Journal of Product & Brand Management, 26(1), pp.26–41. Available at: https://doi.org/10.1108/JPBM-06-2016-1215

[4] Huang, C.-C. (2017). The impacts of brand experiences on brand loyalty: mediators of brand love and trust. Management Decision, 55(5), pp.915–934. Available at: https://dx.doi.org/10.1108/MD-10-2015-0465

[5] Gavilanes, J.M., Flatten, T.C. and Brettel, M. (2018). Content Strategies for Digital Consumer Engagement in Social Networks: Why Advertising Is an Antecedent of Engagement. Journal of Advertising, 47(1), pp.4–23. Available at: https://doi.org/10.1080/00913367.2017.1405751

[6] Schreiner, M., Fischer, T. and Riedl, R. (2019). Impact of content characteristics and emotion on behavioral engagement in social media: literature review and research agenda. Electronic Commerce Research. Available at: https://doi.org/10.1007/s10660-019-09353-8

[7] Woodside, A.G., Sood, S. and Miller, K.E. (2008). When consumers and brands talk: Storytelling theory and research in psychology and marketing. Psychology and Marketing, 25(2), pp.97–145. Available at: https://doi.org/10.1002/mar.20203

?[8] Kashdan, T.B., Stiksma, M.C., Disabato, D.J., McKnight, P.E., Bekier, J., Kaji, J. and Lazarus, R. (2018). The five-dimensional curiosity scale: Capturing the bandwidth of curiosity and identifying four unique subgroups of curious people. Journal of Research in Personality, [online] 73, pp.130–149. Available at: https://www.sciencedirect.com/science/article/abs/pii/S0092656617301149