CONCLUSION

AI-technology is increasingly being implemented in digital marketing communication. Since the technology is continuously being implemented, it was considered interesting to investigate how this implementation might affect the ability of building customer-based brand equity.

A limited amount of research is already done within the area of AI-technology and branding, especially within digital marketing communication. Therefore this study sought to explore the following research question:

How does the application of current artificial intelligence technology in digital marketing communication affect a company’s abilities in building customer-based brand equity?

The findings discovered that AI in itself is used as a buzzword, where using the term alone can be used for marketing purposes. Combined with this is the fact that there is no transparency in the technology. People do not understand the technology and are blindly trusting the results it produces.

As people do not stand the technology, there is a degree of overselling, as it can often be perceived as capable of something it is not.

The term of AI is usually used as an umbrella term for several sub-technologies. However, there is no strict definition of AI, meaning that some people perceive AI as only some of the technologies it involves.

It was however found that when used for branding purposes AI can easily work with quantifiable attributes, as these can be identified and worked with. AI can improve on quantifiable-aspects and comprehend an amount of data that is not possible for the human mind making it very valuable.

When it comes to attributes and aspects of a brand that is not easily quantified AI-technology is struggling with identifying and working with these.

AIs ability to work with such aspects rely on the data-providers ability to define the aspect and

provide a quality data training set.

108

The training data set gives AI the ability to learn what is right and wrong in dealing with a specific task. This makes it difficult for AI to be creative, and therefore where creativity is needed human intervention is necessary.

AI is constantly learning by trial and error, meaning that it will fail, and if it is facilitating digital marketing communication this can damage the brand experience of the individual. Therefore, it is essential to set limits for an AI, as the technology at the moment will be failing massively when dealing with not directly quantifiable tasks, and it is strongly suggested that human supervision is always done to secure the quality of the results, the AI delivers.

As AI will always fail, companies should go slow when implementing the technology as this will give the ability to build up more historical data over time, making it more precise. Here it is

essential to highlight that the harder metrics are to quantify, the more examples must be provided in the form of training data for AI to perform at an acceptable level.

The use of AI-technology can to a high extent contribute to creating a relationship with the customers, as it can profile each customer based on data and personalize the approach to each of these based on the individual. These personalization abilities can increase the customer

experience and contribute to creating brand loyalty.

The findings discovered that the current brand equity models are not applicable when used in an AI-driven digital marketing communication context, making it necessary to construct a conceptual

model that provides a guideline for the technology’s ability of building brand equity.

As a result of this, a conceptual model for building customer-based brand equity in an AI-driven digital marketing communication context was formed based on the findings, which displays to what extent AI-technology can work with and facilitate different branding aspects.

As quantifiability is vital in order for AI-technology to work with specific tasks, this becomes a

crucial factor in AIs ability to work with brand building blocks. The brand salience elements, like

brand name and logo, already have pre-assigned labels when used online, making the elements

identifiable for AI-technology.

109

Therefore this is the easiest brand building block for AI to work with. The brand performance aspect covers, e.g. price and product characteristics, which are easily quantified also making these easy to work with. Brand resonance covers the relationship between customers and the brand. AIs ability to contribute to the enhancement of the relationship between the customer is a strong attribute of implementing the technology in digital marketing communication. This is a result of its ability to benefit from a significant amount of data, and personalizing communication to the individual.

However, the brand building blocks covering more intangible aspects, which are more difficult to quantify, are harder for AI to work with. Brand judgement covers the opinions of customers e.g.

the perceived quality. AIs ability to work with opinions is limited, as the accessibility to data about opinions are limited. However, sentiment analysis’ of text is able to identify opinions to a certain degree. The data input in text can though be limited making it difficult to do a precise estimate on a opinion and require a significant amount of training data. Whether the identified opinion from the sentiment analysis reflects the actual opinion of the customer is though not certain as it is not possible to investigate further. Brand feelings touch upon a customers feelings, but like brand judgement the accessibility to data is limited, and even though estimates can be made on the feelings of a customer, these may not reflect the reality as a result of the data limitation. The toughest brand building block for AI-technology to work with is brand imagery. Brand imagery covers the hedonistic aspects of a brand e.g. values and heritage. This is a result of the difficulty of quantifying this building block. For AI to work with these, it demands a lot of a data-providers ability to define values and heritage, which may be difficult, and can create biases. It will also require an extensive training data set in order for an AI to be trained well enough to get a grasp of these aspects, which may require extensive resources to create

The findings shaping the model highlights that the use of current AI-technologies in digital

marketing communication affects a company’s ability to build customer-based brand equity. This

is a result of the technology currently not being able to facilitate all aspects needed for building

customer-based brand equity. It does, however, show that elements that have historically been

difficult to improve are significantly easier through the implication of AI, e.g. the relationship with

the customer.

110

The study is based on in-depth interviews with a group of AI-experts who have knowledge within digital marketing, through a grounded theory approach.

As the study takes a post-positivist worldview the quality of the result is evaluated based on the validity and reliability of the study. To enhance the validity of the study, detailed descriptions of the data collection and analysis process are included in the study together with transcriptions of the data used for the analysis and the findings that the analysis led to. The inclusion of the

transcripts increases the reliability of the study, together with the names of the interviewees being displayed rather than being anonymous. As transcripts were analyzed independently, the

researchers conducted an intercoder agreement, and the differences were resolved by discussion among the coders. The paradigm does, however, highlight a researcher bias, resulting in the researchers influencing the results of the study. It is however considered replicable based on the provided data and the worldview of the paradigm.

The companies who might benefit from implementing AI-technology is split. Older companies have more access to historical data and more resources to implement the technology, while new

companies might have limited access to historical data and fewer resources to implement the technology. These companies can draw on external technology providers to deliver

AI-technologies, but this causes a dependency on the supplier, which might be costly. The findings also highlighted that what the technology may be capable of in theory may not be how it is used in practice making it even more challenging to distinguish high-quality AI-solutions from low-quality AI solutions.

It is however emphasized, that companies who are differentiating based on hedonistic aspects of a brand can face several problems if implementing AI-technology for their digital marketing

communication, as it is hard or impossible for the technology to take these aspects into

consideration. It is however beneficial for companies differentiating on quantifiable aspects, e.g.

price or delivery time to use AI-technologies for digital marketing communication, as these are

easy to identify and work with for the AI.

111

The extent to which companies should personalize their digital marketing communication is also questioned as it can cause problems if personalization and optimization are not limited. It can be

considered as ‘spooky’ by the customers or even cause problems as it is only able to act on the

data provided, and can therefore not take external factors into account.

Whether AI should personalize company values to the individual should also be taken into

consideration, as it can improve the relationship with the customer, but causes problems in being

consistent in the communication of the brand, e.g. online vs. offline.

112

In document ARTIFICIAL INTELLIGENCE IN DIGITAL MARKETINGS EFFECT ON BRANDING (Sider 107-112)