The role of strategy in the age of AI marketing

December 11, 2017

By Kevin Mason, Strategy Director 

Go to any dinner party where marketing and advertising folks are gathered at the table, and the conversation will inevitably turn to how artificial intelligence (AI) is going to change our profession in the next five years. The current narrative goes something like this:

“AI software will replace the labour-intensive bits of our job – media buying, testing, optimisation, targeting, execution, analytics – and will do these things more efficiently than we ever could. We’ll be left to concentrate on strategy and creative.”

Given where we are in AI evolution, this is about as far as our thinking goes. Let’s face it, very few of us have an AI specialist in our team yet. I’ve been lucky enough to have a demo from one of them, and from our first glance of its capabilities, and the early results coming from case studies, it certainly appears impressive.

I think most of us agree that AI marketing automation is inevitable. But what will ‘concentrating on strategy’ look like in this new reality. Will it be, as promised, all the fun stuff we do today, without the drudgery? I don’t think it will. I think it’ll be very different, and here’s why.

From master and commander

Think about our relationship with marketing technology in the last ten years.

We’ve developed a vast array of tools to execute and optimise our strategies. Our data scientists have calibrated statistical engines to gather more and more insight into our audiences. We’ve programmed our automation systems for mass customisation. We’ve targeted our audience across networks, social platforms, search engines and access devices. We’ve automated our bidding processes to get the biggest bang for our buck.

We’ve been programmers. Masters and commanders of our technology. All in the service of our strategy.

This relationship is just about to fundamentally change.

To the setter of boundaries

In the not too distant future, we’ll feed our AI engine with the data we’ve collected so far.  Our strategy will form its terms of reference and its boundaries. Its starting point.

From there, our AI team member becomes an autonomous marketer, exploring the relationship between its terms of reference and the sea of data it finds in executing its duties. Operating within the boundaries set by our strategy, it will hunt out correlations that help it achieve its objectives. It will do this with speed, and at scale.

Our AI friend’s successes will be of its own making, and we’ll hopefully celebrate by reaping the commercial advantages.

But here’s where I anticipate our strategic role will become markedly different.

Our interest will turn to where the process found dead-ends at the boundaries of our strategy.

Our interest shifts

Let’s say, for example, that we’ve been targeting equity investments to 50-plus-year-old, mid-senior level managers and directors, as a good way to top up their pensions.

Our AI engine discovers that it gets more response from people in this profile whose social feed features pictures of Labradors. As far-fetched as this may sound, it’s exactly the type of correlation it’s able to find and analyse. It looks for more Labrador sharers, and discovers more responders with the same profile down to the age of 40. Below this, responses drop off.

When our AI colleague shares this insight with us, we might conclude that Labradors tend to be owned by families with traditional, conservative values, a good fit for our proposition. Thinking this through further, we might visualise an entrepreneur with a younger family. What type of dog might they own? A French bulldog maybe? They won't fit our profile for the pension proposition, but they may be interested in opportunities for growth and diversification in their investments.

We’ll ask our AI colleague to take these ideas and explore further. It will come back with a new set of results and, inevitably, a raft of tangential, unexpected dead-ends at the edge of our strategic boundaries.

We become mentors

So, in this glimpse of the future interaction with our AI colleagues, our relationship has shifted. Where previously technology provided us with the tools to do our job, now we are mentors to our AI students.

If this is the case – and it doesn’t take too much of an imaginative leap to conclude it might be – then the most prized AI strategists will be those who develop the deepest relationship with their AI students.  

Ultimately our role will be to teach our AI friends about the human condition.

Things are about to get interesting.

We’re planning some AI events and trials in the new year. To join our invite list, send your email address to