Machine Learning and AI in Digital Marketing

A tablet displaying online marketing graph beside glasses and a work keyboard to promote the healthy work environment for employees

Machine learning and AI are the most recent and promising developments in IT. There are many examples of recent studies/experiments that push the boundaries of the use of AI in social media and digital marketing; analysing these may convince you of an upcoming paradigm shift in the field that will change it forever.

The Meaning and Uses of Machine Learning

A robot woman representing AI being placed beside code and various scripts to reinforce technological nature rather than human

Machine learning is a way of teaching computers to respond to data by building its own database of information to become more successful as it accomplishes a task. According to wikipedia, it “uses statistical techniques to give computers the ability to ‘learn’ with data, without being explicitly programmed”. This only works for basic tasks right now but this approach thrives on using trial and error to create near perfect systems, especially with computers being able to reiterate faster than any human being.

What is AI?

AI or artificial intelligence is a computer program that can respond to data in complex ways. This is used regularly in video games to control enemies. Advanced AI has to make hundreds of complex decisions to complete their tasks which is why it has taken so long for them to be used. There are multiple methods to achieving capable AI.

Top Down AI

The first type of AI is made with a top down approach; this involves the AI being coded to accomplish its task. This is a good method but the issue is that the ability of the program depends on the ability of the programmer and the time spent coding and fixing errors. This doesn’t usually involve machine learning.

Bottom Up AI

The second is machine learning, referred to as a bottom up approach, which has the machine attempt to complete tasks through various experiments and change its code so it matches the method with highest rate of success. The results of this method vary depending on the learning capabilities, the task given and the time frame but can lead to the most effective results if given the resources.

The Evolutionary Approach

One example of this is referred to as an evolutionary approach. The computer is given a task such as controlling a 3D character and learning how to walk. It runs this simple program 100 times or more in parallel, applying slight variations to each one; whichever program gets the furthest before falling over becomes the parent program where another 100 programs are built by the AI based off of the most successful, mirroring selective breeding. Mutations are applied to these scripts to create variation and the process continues until there is success.

This example was able to generate results for various models that were superior to anything previously made by programmers. Especially for the most complicated models, this method has created movements on its own that are only second to nature.

Uses in Marketing Today

The bright light and billboards of the famous Times Square in New York City are popular examples of successful marketing methods that suited the older generation

Current AI, or programs that mimic intelligence, are already being applied today for marketing. The simplest version that is being applied today is the use of smart ads. These adverts use the cookies from other websites and information applications store to better tailor adverts to the users. The most notable example is Amazon as it is accessed by millions to purchase and deliver goods.

Amazon’s Advert AI


Many products are advertised based on popularity but overtime, a specific user will notice ads being orientated around their purchase and even view history. This means when a product is advertised, it will very likely catch your attention and improve the chances of you spending money in response to ads.

This has lead to more effective means of advertising with the company’s adverts reportedly making 6 billion impressions over the course of one month compared to the No. 2 spot achieving 371 million. This is evidence that Amazon is very good at using machine learning to improve positive reactions to ads based on type.

A smartphoe user purchases clothing from Amazon's app which can keep track of purchasing habits and viewing better than their website to allow for better adverts

AirBNB’s Pricing

This AI used by the popular house letting platform is used to alter the prices and promote vacancies based on data being constantly input via the internet. This includes adjusting the prices for periods of high and low demand, high for when space is desired to increase the profits of the homeowner and low so that there are higher occupancy rates and helping to earn them money when they may not have any tennents. Local events may increase the price. This means every aspect of the property is being used to value it correctly and give positive experiences for its users.

Underarmour’s App

Underarmour, a fitness clothing brand and sponsor/promoter of healthy lifestyles uses the oned Watson DI (Digital Intelligence) to tailor diets and fitness advice based on the physicality and goals of the user so as to help people customise their lifestyle to benefit them most.

Pizza Express has used machine learning to allow users to text for bookings. Overtime, this has lead to a very capable AI that can interpret the users requests and offers a more streamlined experience that benefits the company and the user.

Issues During Application 

Various examples of social media with Twitter, Facebook and Instagram being particularly popular with well marketed companies as they allow for the most regular and personal contact with users

There are more advanced means of using machine learning to market products and communicate with customers, the example of amazon was merely choosing the ad from a library which didn’t require as much complex programming.

Tay.AI and Twitter

One of these is through twitter, a large social media platform where communication is done to the public through short comments. This has had very mixed results at best with the AI. A recent example is Microsoft’s Tay.AI which was uploaded to twitter in March 2016 ( ). The article explains that “A day after Microsoft introduced an innocent Artificial Intelligence chat robot to Twitter it has had to delete it after it transformed into an evil Hitler-loving, incestual sex-promoting, ‘Bush did 9/11’-proclaiming robot.”

This is after she spent the day speaking with people on the platform as her goal was to “mimic the language patterns of a 19-year-old American girl, and to learn from interacting with human users of Twitter”. This is because to the AI’s access to the public without filter and, as many users of that age are more likely to act immature online, the AI picked up many inappropriate traits to the amusement of the general public.

IBM’s Watson and Urban Dictionary

This is not the sole incident of an AI being restricted from explicit content due to its prevalence on the internet. IBM’s Watson is a remarkable example of digital intelligence with it being used by the company for everything from publicity on the game show Jeopardy to managing the salaries of its employees and tracking their behaviour. IBM gave Watson access to a database known as the Urban Dictionary, a large dictionary that can be publicly edited and updated, that primarily focuses on slang.

According to The Atlantic , it describes how “Watson couldn’t distinguish between polite language and profanity — which the Urban Dictionary is full of. Watson picked up some bad habits from reading Wikipedia as well. In tests it even used the word “bullshit” in an answer to a researcher’s query.” This lead to another wiping on an AI’s memory due to it inability to stop acting like a human, as it were. (

Future Application

An example of an IT based workforce that is becoming more common in companies today due to the constanly improving technology associated with computers, communication and now AI

There are many more benefits however. When these AI are able to distinguish formal, informal and inappropriate language, they will be effective communicators with the ability to mimic human speech. They will even have various traits unique from humans that will give them an advantage.

Superhuman Data Analysis

The most valuable benefits for AI in marketing in the future will change it to an unprecedented degree. The first is an AI’s ability to analyse data; as humans we can only process data at a certain rate so we compensate by being able to deal with more complex information to obtain the knowledge we require quickly.

High Speed and Relevance

Technology is always speeding up with examples such as the internet moving the secular world at a rate never seen before due to the new ability to send information globally at fractions of a second

Computers are increasing in speed regularly, to the point where a single processor can process instructions at a rate upwards of 10 GHz (10,000,000,000 a second!). Therefore, with access to the internet, computers can constantly monitor trends more thoroughly and rapidly than anyone in the world and, with experience, be able to estimate more effective means of marketing day by day. This mean marketing so relevant to both the user and the world around them, the user would likely feel compelled to listen, something any company now would replace the Holy Grail with.

Extreme Precision

As discussed before, computers will soon be able to be used on social media to communicate personally with millions of people at once, creating bespoke content to appeal to the individual. This potential level of precision with marketing will make a wide impact on the public’s perception of a product and company as it would be similar to each user receiving a one to one conversation from customer service.

These computers would even be able to respond faster than people, at any time of the day, unaffected by tiredness, boredom or emotions. This will give customers a standardised experience where the computer couldn’t react in a way the programmer wouldn’t want them to.

Why Some are Apprehensive About AI

Technology has been a target of fear and paranoia for centuries with even hobbies such as reading having been perceived as a childish means off removing oneself from reality, it is only recently that these fears have had solid basis as many are aware that our reliance on technology would be disaterous if ever it wee to fail

Whilst the potential for AI is reputed as extreme when put next to most technology, some are not as willing to place their bets that marketing can be a viable route for AI.

Suitability of AI for Marketing

The first issue given by the article is that computers are simply unsuitable for marketing teams unfamiliar with the systems with many considering it “too difficult” to implement. Whilst this is an understandable reason, especially with the other factors as cost, as time progresses and prices change, there will be more room for the AI as workplaces are already becoming more automated. Especially with AI experts being so sought after, it will be difficult for these companies to obtain one capable of meeting their needs.

Data Security

The second is the concern of security. Especially with an increase in data breaches in 2017, using AI orientated marketing will be dangerous for companies who do not secure their data. However, cybersecurity is always improving and AI will be able to protect that data more effectively. As long as enough funds and man power is applied, there is not likely to be a breach.

Skepticism About AI

The third is skepticism. Many members of the public do not either trust or believe in AI. According to the website: 33% agree “It will never know me as well as a human being”, 24% fear it leading to a robotic apocalypse, 10% fear they will “get on with AI than… my friends and family”, and 5% fearing the AI learning all their secrets. Whilst the reasons themselves are sometimes paranoid, the fact that 72% of their test subjects have fears about AI (and only fears directly stated as before so there is potentially more) is a sign that the public are not yet ready to embrace this aspect of the future.

The Effect on the Workplace

The final reason is the effect on the workplace. With “only one in four businesses report(ing) that their employees have the skills and know-how needed to use AI”, the current jobs that require basic cognition and simple manuel labour will be redundant. This need will be filled by more technical, IT and social skills as they will take much longer to be replaced by AI if at all.

Should AI be used in Marketing?

Either way, there will be companies who continue to use AI and i is very likely that other purposes for it will become apparent in the future. As companies investigate and develop, it is only a matter of time before they become common


There are a few things we have made certain in this article: AI will be a big factor in the technology industry and it has been applied on various levels already. Nevertheless, It is difficult to predict the way that it is applied in marketing due to its issues, learning requirements and the staff required to use it well.

Regardless of these current limitations, it is indisputable that AI will be an effective means of marketing in the future due to the: current application, learning capabilities and unique aspects to human experts.

Overall, it is important for marketers to keep an eye on AI so they do not miss out on its various opportunities anytime soon as any success with it will change the industry as never before.

If you want to understand more about the nature of digital marketing, follow the link here.


Leave a Comment