It seems that every year is touted as the year of Artificial Intelligence (AI). Yet year after year we fail to see that claim come to fruition. Where is this inundation of smart technology in our day to day lives? Is this a classic example of hype and media attention getting the better of common sense?
It is difficult to fully establish the state of play with Lyre Birds occupying the ecosystem giving a false sense of the extent of AI saturation, but we will try all the same.
So, what is AI?
Before we continue, let’s stop and get a few definitions straight. Automation is the use of software capable of doing tasks or processes automatically. It is not AI. Machine learning is also not the same thing as AI, even though nowadays these terms are often used interchangeably.
Traditionally, AI is the broad concept of machines being able to carry out tasks in a way that we would consider smart. Machine Learning Technology (MLT) is an application of AI whereby you give a machine access to data and allow it to learn for itself. This learning element is crucial and, for all intents and purposes, is the key to something possessing true intelligence.
Is it all hype?
I believe there are areas where the impact of AI goes beyond uncommercial hype.
Search has come a long way since the early days. With a combination of natural language processing and machine learning, search engines are grasping our intent in ways we could not have imagined 10 years ago. It is vital that marketers grasp and understand these changes as fast as possible.
Programmatic buying and selling of ad space has made a huge difference to the reach and targeting capabilities of online advertising, but under this definition, it is merely automation. However, the application of propensity models off the back of machine learning algorithms will revolutionise how you target ads at your most relevant customers. Predictive analytics is likely to be another pivotal area. The use case is simple, one would have machine learning technology applied to help predict the impact of an ad or placement before launch, saving time and resources from dead-end investment.
Imagine this scenario: a customer walks into a mall, and cameras instantaneously detect their fashion, sense their friends, even their facial cues and body language. This information is then pushed through MLT and compared with masses of data collected about consumers and their shopping tendencies. Your propensity models then indicate that this person is worth the media spend. Ads are then shown to this person as they walk through the mall: if they walk into your clients shop and buy the items you have so thoughtfully picked out for them, you are awarded the conversion. If not, then at least your AI will learn.
What do these examples have in common?
Data is all important!
Amassing enough reliable data to implement is going to be the biggest hurdle to widespread application, especially with the growing data restrictions.
It needs to be up to date, processes need to be in place to handle it, it cannot be siloed, and you need lots of it. You can have the most advanced machine learning algorithms but apply it to crap data sets, which plague our industry, and the output is worse than an old-fashioned human.
How do we need to influence change?
AI is expected to have a revolutionary impact on the advertising industry. However, it is going to need huge investment to get the computing power, software, developers, and algorithms in place to realise that potential.
For the most part, it won’t be the advertisers and agencies leading the charge from a tech point of view. But what should we be doing to prepare for this world?
The role of the Campaign Manager
One of the biggest impact areas for biddable media will be understanding and reacting to the AI advances. In terms of campaign management, this means remaining on top of industry developments, analysing how ads function in changing environments and understanding how this will influence media strategy.
The role of the Leaders
Without the direct technical knowledge, decision-makers will have to focus on balancing three variables: how potentially valuable the technology could be, how mature the technology is, and the ability of the organisation to tolerate and manage risk.
Traditionally when thinking of outsourcing, thoughts turn to cheap labour abroad. Going forward agencies will be outsourcing to intelligent platforms and services. Deciphering between those that will benefit your company’s top line and those that are being misappropriated will be the key in developing your tools.
Final thoughts
There is such huge potential for AI in our industry, however, it is abundantly clear that we have developed ‘technology hype syndrome’. We have all identified the huge part it must play without fully understanding the infrastructural shifts needed to realise widespread adoption. We have convinced ourselves that we are further along the road to intelligent machines than we really are. For agencies and advertisers without access to data powerhouses, adapting to shifts and tool selection are going to be the keys to revolution.
Darwin said that “It is not the strongest that survives, nor the most intelligent. It is the one most adaptable to change”. Seemingly there are still lessons to take from this phrase, almost 200 years on, in the competitive environment of advertising.
