Real AI Uses: Marketing | AI For Business #2

When people talk about AI in marketing, they often mention big words like ‘big data analytics’, ‘leverage customer data’, ‘revolutionize customer acquisition’.

There is much to say about the benefits of using AI in marketing - but you already know that. You have probably already read a dozen articles touting AI as this omnipresent, all-powerful ‘thing’ capable of many, many great things, one could even say it's capable of everything!?

Then you probably come back from the clouds and realise that you have no idea how exactly AI can help you right now.?

After all this grand talk of AI being the lord and saviour of marketing, marketers often feel like AI is out of their reach, that their needs and wants are far below the realm of AI, and it is either used by?

1. Huge corporations like Google just because they can,
2. The largest marketing agencies that take their data extremely seriously and perform all sorts of complex data analysis.

And they couldn’t be more wrong.?

In reality, AI can - and has already - find uses in almost every daily activity a regular marketer performs. AI is REALLY used by

1. Social media managers to automate caption generation
2. Content marketers to generate articles and web pages?
3. Customer care managers for review analysis
4. Marketing managers to quickly analyse competitors and what people say about them

All these ‘mundane’ tasks being automated by AI is the actual reality of tech innovation for most marketers.

I speak from experience. I’ve been a part of multiple projects aimed at making life easier for marketers through the use of AI and have seen how successful these projects can be.

Here’s a list of several ways AI is used for marketing - and not in a pretentious way.

1. Caption generation for Instagram

Scaling content generation is important for marketing agencies who have many clients, each with their own social media.

Post frequency and level of engagement are two of the most relevant factors when it comes to the success of the marketing campaign on any social media.

Automatic generation of captions solves both of these problems. The texts are generated within seconds, while the texts themselves are created to maximise engagement.

An AI-based system takes an image as input. First, it processes the image to extract objects and classify them.

This helps the system to understand what is going on in the picture and how best to describe it.

Each object has a confidence score. Only high-confidence objects are used for caption generation.

Next, comes text generation. Have you ever seen those odd tweets generated by AI??

Yeah, the goal is to do NOT that. The goal is to generate text which would be indistinguishable from that written by a human.

This is achieved by training a machine learning model using a collection of Instagram captions written by real users, hashtags included.

After the ML model has learned how humans caption their images, it combines this knowledge with high-confidence objects and generates human-like text:

For the marketer, the process is easy: upload an image, receive a caption, enjoy.

2. Aspect extraction from product reviews

Product reviews on popular marketplaces like Amazon are a source of customer pains - which a marketer can solve through a carefully written product description.

With high competition within online marketplaces, it is important the products you are trying to sell have descriptions that really speak to the customer.

No to generic ‘this is the most comfortable pair of pants’, yes to personalised ‘the seams are reinforced so these pants definitely will not tear’.

What’s better than reading through thousands of reviews looking for what the customers often complain about? Making AI do it for you.

I’ve been a part of a project where a marketing agency has automated their customer sentiment analysis efforts by implementing an AI-based Amazon review system.

The system processes reviews for any given product and extracts the main points users mention, like a faulty wrist strap on a watch or a good battery life of a phone.

It also ignores any comments about customer service or a marketplace itself, as well as personal opinions like ‘the worst product ever!’ or ‘I love it’ as they are not relevant for marketing content.

These extracted sentiments can then be used to create a compelling product description or…

?2.1 Review-based marketing pamphlets

To push the capabilities of AI even further, the extracted customer sentiments can be used to generate entire marketing pamphlets from scratch.

We at Businessware Technologies LINK have developed a marketing content generation system which analyses reviews of competitor’s products as well as web pages containing description of the product in question.

The system then produces a marketing text based on real customer opinion, detailing top advantages and disadvantages of competitors’ products, describing the product in question on the basis of customer sentiment.

This information can then be used as the basis for any marketing content or as marketing content in itself in the form of a script for a promotional video or a social media post.

3. Video game analysis for market success prediction

Before a marketing campaign starts, one should analyse the product or services that need to be promoted and determine if it's worth the time and effort.

One of the more curious cases I had the opportunity to work with was a mobile video game analysis system which predicts how popular a mobile game will be, along with giving recommendations on what can be changed to improve chances of success on the market.

In the world of mobile games, the speed with which new releases come out and older ones die is incredibly fast. Short attention spans of mobile gamers make them a tough crowd to work with - they are quick to get into a new game but rarely visit it again after playing it once.

There are hundreds of mobile games available for download, the competition is steep, meaning investing resources in promoting a game which is not appealing to mobile gamers can become a huge money drain.

In order to avoid that, a production and marketing agency employed AI to perform digital media analysis and determine how successful the game might be.

The system takes many factors into consideration, like the overall color palette, camera movement speed, frame rate, the frequency of scene changes, etc.

Based on historical data, the system compares the game to other successful mobile games and determines how successful it will be.

Is AI the future of Marketing?

No, AI is not the future of marketing, it is its present. AI has already found a way into marketing activities, and in most cases in much more ‘mundane’ ways than we are used to hearing about.

AI is not this super futuristic unattainable thing anymore. Many marketing agencies already use it to make marketers’ lifes easier, and their marketing efforts more efficient.

I expect to see the interest for AI grow more and more in the coming years, and not from high-tech cutting-edge businesses, but from areas of business who have historically been more conservative when it comes to technology.