19 - ARTIFICIAL INTELLIGENCE AND INFORMATION

Question 1

Artificial intelligence (AI)

Answer 1

Intelligence “demonstrated” by machines,

as opposed to the natural intelligence “displayed” by humans and animals

Question 2

AI can be used in three different ways in marketing research

Answer 2

1. Data collection
2. Market analysis
3. Understanding the customer and “developing key insights”

Question 3

AI for data collection

Answer 3

Customer intelligence, product usage and consumption experience, unstructured marketing activity data, in-car sensors, and retail technologies

• customer intelligence, including data about consumers, their activities, and their environments, can be collected if they use connected devices (e.g., wearable devices, smart fridges, and digital assistants such as Siri)
• product usage and consumption experience can be visualized with Internet of Things (IoT)
• various advanced technologies and analytics can capture unstructured marketing activity data
• in-car sensors can track driving behavior for determining insurance premiums
• retail technologies, such as heat maps, video surveillance, and beacons, can be used for profiling and recognizing retail shoppers

Question 4

The data collection capability of AI

Answer 4

(observable behavioral, survey or experimental)
is not limited to observable behavioral data; it can also be used to facilitate survey or experimental data collection to capture consumer psychographics, opinions, and attitudes.

Question 5

AI and Data Analysis

Answer 5

1. “Automated text” analysis - consumers’ behaviour insights
2. “Machine learning” algorithms and “lexicon-based” text
3. “Big data” marketing analytics

• Automated text analysis is applied to gain insights into consumers’ behaviour (e.g., by analysing customer reviews on a website)
• Machine learning algorithms and lexicon-based text classification is used to analyse various social media datasets
• Big data marketing analytics is a mainstream approach for generating marketing insights. Specific applications include mapping market structures for large retail assortments using a neural network language model, by analyzing the co-occurrences of products in shopping baskets, detecting copycat mobile apps using a machine learning copycat-detection method, and aiding social media content engineering by employing natural language processing algorithms that discover the associations between social media marketing content and user engagement

Question 6

AI and Improved Understanding of the Customer

Answer 6

AI can be used to understand “existing and potential” customer “needs and wants”,

for example, who they are, what they want, and what their current solutions are

Question 7

The major distinction between:
market analysis
customer understanding

Answer 7

customer understanding: involves emotional data about customer “sentiments, feelings, preferences, and attitudes”

Question 8

Example
AI and Improved Understanding of the Customer

Answer 8

“Affectiva” partnered with “Ford” to create “AutoEmotive” “sentiment analysis”, to try to figure out “drivers’ emotional states”.

For potential customers, marketers can use feeling AI to understand what they want and why they are happy with competitors or outside options.

Unilever synthesized insights through social listening and CRM, and discovered a link between ice cream and breakfast. Unilever discovered that there were at least 50 songs with ‘ice cream for breakfast’ in their lyrics (which people were listening to), as well as Dunkin Donuts selling ice cream during the morning part. With this insight, Unilever developed a new brand Ben & Jerry’s with a range of cereal-flavored ice-creams for breakfast.

Question 9

Limitations of Artificial Intelligence

Answer 9

• Using AI requires financial and human resources.

The initial cost of setting up an AI infrastructure is always “resource-intensive”.

Thus, it may not be feasible for all organisations. Investing in AI can mean heavy expenditures on “data acquisition, computing, and storage equipment”,
as well as spending on “recruiting relevant personnel and training them”.

• As with all investments, the “returns may not be immediate”. It may take a while to realise the benefits of using AI.

Question 10

Unrealistic Expectations from AI/Threatened by AI
(feared job losses - bias towards the use of the machine)

Answer 10

Managers need to analyse what is feasible for achieving specific objectives.

The author of this book has worked in the area of social robotics. Deploying a social robot within a business environment often posed challenges. Either the expectations of the business managers were too high for the social robot (e.g., “will it be able to answer the phone?”) or – in some cases – it did not help with their overall communication objectives.

• In another few cases, employees were threatened with the “arrival of a social robot as they feared job losses”.
• In many instances, the “overall evaluation” of the technology was undertaken by individuals who may have had a “bias towards the use of the machine”.