3 practical uses of AI in digital marketing

OPINION / 27th November 2018

There’s a lot of hype surrounding AI and what it can or can’t do. Far from being the dystopian vision portrayed by the likes of Netflix’s boxset ‘Black Mirror’, AI is not a sentient machine and nor will it ever be. AI is a technology that enables machines to sense information, classify it, analyse it and make inferences based upon it. It can predict, act and learn.

Two of its most important characteristics are that it is largely autonomous, meaning that it is making decisions on its own; and it is constantly learning from data and past experience.

Businesses are increasingly able to utilise the power of AI due to the vast amounts of data being generated each day. IBM has said that 90% of the world’s data has been created in the last two years. Yet, with the introduction of sensors and IoT, this figure is dropping to a matter of months.

Machine learning algorithms, or what we think of as AI, are being fuelled by vast data sets and faster computing power and processing. Companies can now process massive amounts of data at a much lower cost than ever before.

As data has become more readily available, and we’re processing it at faster speeds, this has led to rapid advancements in research and improvements in algorithms that we’re using on a daily basis – think Google search, predictive text and everything you see on your Facebook or Twitter feeds, all of which are being underpinned by AI.

So how can marketers take advantage of this new technology? Initially, there are three possible ways:

Analytics

Data analytics is becoming increasingly sophisticated as we now have the ability to use past data to predict future outcomes. Many SaaS companies have built AI into their analytic packages so that marketers can use data to find the most efficient way to target an ad or to personalise a campaign.

AI can take what you already know about your customers, products, services, history, along with macroeconomic trends, and other signals from the market, and predict what might be the best way – for example to create content for a specific audience.

Computer Vision

AI is increasingly being used to understand the contents of a photograph or video, via facial and image recognition. It can determine how many people are in the image, what the weather’s like, whether the scene is set in the countryside or city, beach or mountains.  

But more importantly, the software can detect sentiment, such as whether people are having fun or not. This can help marketers to gain a deeper understanding and provide insights into how their customers are interacting with their brand, which can help inform the next campaign.

Natural language

Natural language enables machines to recognise speech and turn it into meaningful content. It can understand and draw conclusions from text in a news story, or from monitoring a customer service chat log; it can even understand what people are posting on social media or the topics they are blogging about.

This technology is propelling us into a new era of communication. We are beginning to move away from the first iteration of the internet, which is largely URL or app based, where users can go to a webpage or an app and type something into search, and see the content they want to be displayed. As a result of developments in natural language, we are entering a new phase of computing where consumers interact with companies and services using speech, text, chat and, potentially in the future, gestures.

View from ORM

Before adopting any new technology, organisations need to consider their wider strategy and business objectives and then identify areas that could benefit from more intelligent technology. Perhaps AI can be used to improve customer churn or understand the drivers behind customer behaviours, or maybe it can be used to figure out customer attitudes about a particular product or service; or ways of improving the customer experience, or finding out the barriers to customer satisfaction. Whatever the area is that needs improving, businesses need to unlock their datasets in order for AI to be effective.

Neil Clayton Neil Clayton Head of Sales & Marketing Neil Clayton