AI is quickly becoming the new buzzword in advertising, but there is more to AI than just machine learning, the most talked about aspect in the AI family. Writing exclusively for ExchangeWire, Gregory Yates (pictured below), CMO, RICG, explains the basics of AI, the importance of neuroscience, and how marketers are starting to use AI to bring the future into the reality of today’s advertising.
AI – everyone’s talking about it, but what is it, really? There are a slew of companies out there right now that are jumping on the latest trend train with chatbots, machine-learning capabilities, analytics, and so on; but this can actually backfire and pose a problem because AI is an umbrella term with several components (and subcomponents) that can be leveraged in many different ways. Sure, it can help advertisers make a bigger impact, stay competitive, and reach the right audiences, yet there’s still a big question mark when advertisers look at AI technology. Marketers need to first educate themselves on how it could benefit – or not benefit – their company and their customers.
At the highest level, the power of AI lies in this ability to combine emotions and behaviour, overlaid by data, to create personalised experiences delivered to you at the right place, at the right moment, at the right time, with the right device. Remember the movie Minority Report with Tom Cruise? There’s a scene where he’s stressed out and walking by some ‘living’ ads in a mall that read his emotions and try to convince him he needs a beer, a vacation, etc. Here’s the thing – this is actually becoming a reality and there’s a huge opportunity for advertisers to up their AI game.
AI & Neuroscience 101: Understanding the Basics
I want to start by first breaking down the core concepts and components of AI and Neuroscience. The basics can demonstrate the true power and promise of these combined technologies before taking the first step down the AI path. Here are three ways that I break down AI, Neuroscience, and the combination and use of both:
Breaking down AI
In its most basic form, AI can be described as methods to reproduce what’s going on in the human mind. What’s interesting is that 99.9% of AI is determined by eliminating data and/or making assumptions, i.e. regardless of everything else, let’s just come to a conclusion based on a smaller set of things we know. Let’s take a quick look under the AI hood to see the different parts that can be laid out in an almost hierarchy-type fashion.
Machine Learning: This utilises algorithms to process data and takes those leanings and applies them to like data sets of probabilities of what it thinks is the next outcome (e.g. customer service chatbots for airlines that are predicting future desired flight deals).
Cognitive: This builds on machine learning, but infuses a ‘human mind approach’.
Deep Learning: This takes cognitive to next level and uses neural networks – the brain is composed of neural networks and now computers try to achieve that by simulating this process as though they’re human brains processing the data.
Analytics: This is AI’s most basic form and entails the simple processing of data, there are three most used types: predictive, statistical, and deterministic.
Importance of Neuroscience
According to human behavioural studies, an individual can make up their mind up to 10 seconds before they even realise. Neuroscience relates to this emotional component; so where AI can help you understand behaviour, neuroscience taps into one’s emotions so you can engage with them on that level and influence their decision-making process.
Neurometrics: The two tools for this are electroencephalogram (EEG), which measures the brain’s electrical activity, and functional magnetic resonance imaging (fMRI), which measures brain activity by observing changes in blood flow.
Biometrics: These are things that are ‘of the body’, such as facial expressions. The Facial Action Coding System (FACS) is a good example where technology can be used to understand predefined microexpressions of the face. Galvanic Skin Response (GSR) is another helpful biometric which is based on reading your sweat (electrodermal activity) to gauge emotional arousal. And there is also eye tracking, which observes a person’s gaze, fixation, and other eye movements to understand their focus. If you triangulate this data you can truly understand how a person is feeling.
Psychometrics: These are based more on a qualitative Q&A method with testing on a 1:1 basis to gain more colour surrounding the experiment, including recall and other verbal conscious dialogue.
AI, Neuroscience, and Advertising
So, now that we have the nitty gritty scientific items out of the way, let’s look at some examples where AI – or at least parts of it – are being leveraged to help advertisers and brands capitalise on this latest tech craze.
Let’s first look at Chatbots, which are designed to simulate a conversation with another human. Brands across several industries are developing Chatbots, or just bots, that leverage popular messaging platforms like Facebook Messenger, iMessage, and Kik. The power of bots is in the natural-language processing actions that allow brands to better understand you, learn your preferences, and deliver targeted ads that you’ll be interested in. Service-oriented brands are well-positioned to benefit, because bots are becoming important tools for consumer discovery. I recently heard of a company called Forkable, a ‘lunch bot’, that learns what people like to eat and then delivers a different lunch every day to their office. From an advertising perspective, specific campaigns on a bot like this can interact and engage with individual users in a more authentic way based on all of the data it has collected on them over time. So, while bots are definitely still in the nascent phase, they’re starting to demonstrate their value to the advertising community.
When we dig into the neuroscience side of things, it opens a whole new world of possibilities for advertisers. Not only to make creative and relevant ads, but be able to tap into human emotions and subconscious reactions to inform their strategies, so there’s literally no guess work or trial and error that needs to take place. Look at what Facebook is doing with AI – they’re building neuroscience labs that use things like EEG and biometrics to get a deeper understanding of how people react to things like watching shows or scrolling through their news feeds. Their goal is to capitalise on the power of neuroscience by helping advertisers, brands, and publishers maximise the impact of their content across different devices and platforms.
Another great example is a test that was conducted by Nielsen where they looked at 100 ads across 25 brands and tested individuals’ reactions to them by using EEG. The goal was to understand above-average versus below-average sentiment and reactions. They found that the ads that had below-average emotional response decreased a brand’s sales by 16%, whereas above-average responses increased sales by 23% – this is a powerful tool for advertisers.
This is all great, but how do I even get started?
The above examples show us that brands are indeed trying to tap into AI to help boost their advertising initiatives, but there’s potential to do so much more. AI can ultimately make those million-dollar advertisements successful and drive sales. Here’s are three things that I tell people about when they’re starting to think about dipping their toes into the AI pool:
Identify the objectives. Focus on what problem you’re trying to solve. You need to build a case to show the need for AI, as it relates to your business objectives and advertising initiatives.
Don’t be afraid. Brands that are willing to grow and drive adoption are winning, whereas brands that have no interest are falling by the wayside – in fact, 80% of AI-adopters replacing their roles will retain and retrain employees, so it isn’t something to be afraid of at all.
Do your research. There’s no one-size-fits-all approach to AI, but start by looking at data you currently have available in-house and leverage it with an AI company to process to understand how it could potentially benefit your advertising strategies. Data can be something as simple as an excel spreadsheet. This will help you understand the insights you can get from AI before taking the next step. It will also help you understand the quality of data and if it should be better structured, formatted, and if there’s anything missing.