Whether you work in media or not, artificial intelligence is the trend that closed 2016 and will dominate 2017.
Just look at the TV shows wowing the critics lately. On this side of the pond, we have Humans, a vision of the near-future where androids reach sentience and fight for recognition, equality and, increasingly, their "lives".
Or stateside, where we have Westworld – a rich man’s virtual playground, populated by robots with decision-making facilities designed to do humans’ bidding with unpredictable and bloody results.
Fortunately, both examples remain firmly rooted in fiction and will hopefully stay that way.
Today’s AI is an altogether more practical affair, especially in adland.
Anything you can’t do 55,000 times, AI can do for you
First of all, as marketers, we need to stop leaping too far into the future. We are not teaching AI to take over from us and AI isn’t going to teach itself that either.
Increasingly, we are teaching machines to "learn" but, for the time being, AI is a very effective robot. It can learn to do repeatable, programmable actions that humans also do but AI can do it faster and more efficiently, with fewer mistakes and greater impartiality.
Therefore, in our fast-moving media environment with so many moving parts, so much decision-making and so little time to do it in, AI can be a real boon. A media plan requires roughly 55,000 decisions to be made – making it easy to see why AI is becoming an attractive option for taking over that heavy lifting.
Additionally, AI can only be "taught" to operate in a single area. It can’t generalise. As an analogy, AI could choose the paint it needs to paint a picture but it couldn’t decide what picture to paint.
It is why we create bespoke algorithms for our clients. So while it’s a new and shiny tool in the marketer’s toolbox, it’s still going to need bright minds to point it in the right direction.
The science bit
AI isn’t going to totally replace human input in advertising. Our unique ability is to strategise, future-gaze, anticipate the ebb and flow of consumer behaviour and deal with the capriciousness of the customer. This feeds creativity and is the engine behind successful campaigns.
But brands aren’t successful on creative juices alone. Brands that win are also efficient – making the most of resources, planning intelligently, rigorously and sometimes ruthlessly. And those are conditions that are ideal for AI.
AI is like a juggler with 12 arms. The juggler isn’t questioning the colour of the balls, they just know how fast they can throw them, how high and how many can be kept in the air at any one time. And a juggler with 12 arms is a damn sight more impressive than one with two.
"Brands aren’t successful on creative juices alone. Brands that win are also efficient – making the most of resources, planning intelligently, rigorously and sometimes ruthlessly. And those are conditions that are ideal for AI"
And AI can crunch the numbers. With the right coding (a machine is only as good as its code, even so-called intelligent ones), AI can extrapolate all sorts of scenarios from the endless flow of data that threatens to drown companies at any moment.
It has the time and energy to look for those hidden details that provide the vital missing link in customer journeys that transform a media plan from almost there to stellar.
Always on and always churning out new insights, AI is the engine behind continual testing and real-time learning that finds those nuggets humans can then take and apply creativity to.
Using AI to deal with emotion
Furthermore, AI’s current inability to "be" emotional means it is ideally placed to take media from being an art and draw it closer to being a science. This in turn gives the chief marketing officer greater clout in the boardroom as they confidently speak the language of the rest of the C-suite. More black and white, less shades of grey.
An AI-enabled system can only find the most appropriate media mix. It can only advise on the best media buys based on the best combination of statistics and metrics.
AI can also be used to crunch the numbers on viewability or reach, projected ROI and optimal frequency. We use it to feed in all the disparate strands of data into a predictive modelling platform that can show the impact of any given media strategy. And, through that, we can give concrete evidence as to what has been done brilliantly and what kind of success those learnings will drive in the future.
You could argue that an AI system is only as good as the inputs – and that really is the case. Scepticism around the validity of some media platforms’ viewability stats is real – and if false figures are provided to the buy-side AI, of course you could expect a biased output.
But what if the media platform’s view-ability figures are also provided by AI? There is no room for interpretation. If an acceptable standard is agreed of one, two or however many seconds of video, sound or no sound, then all AI operates on the same standard. There is no question of favouritism or massaging the results.
The industry is crying out for this kind of impartiality and accuracy, and 2017 is going to see some dramatic increases in demand for AI. It’s going to be the only way companies can keep up with the proliferation of channels, platforms and consumer trends.
Is AI replacing data scientists, analysts, media planners, buyers or creatives? Not entirely. AI is only a tool. A fast-acting, wide-reaching multi-tool. It still needs the emotion, gut instinct and unpredictability that only humans can provide. But AI offers a high-speed, high-accuracy platform for them to work on.
Carl Erik Kjærsgaard is the chief executive and co-founder of Blackwood Seven