So, Alex, what’s your quick take on where AI/ML stand for marketers?
Well, against a background of a world that’s being increasingly automated—faster than we might have predicted five or ten years ago—automation can be both exciting and pretty scary for people. It can feel like it’s taking away control. But the truth is that machines are very good at doing some of the things that we humans are not, and we need to embrace that.
What are people so scared of?
I think many assume that automation is complicated, with high math and a lot of expense. And we are also unsure about what it is exactly that machines will and won’t take over. But the good news is that they’re taking on the tasks that humans weren’t built for: computing billions of decisions every second, tailoring content to every single person in a customer base that could count in the millions, and so on. They never tire. Never fatigue. Never have a bad mood.
How can marketers approach automation differently?
As ever, simple is the way to go—what we are calling a trust, test, and tune approach. First give the machine a simple measurable goal, build a deep understanding of how it’s working out the solution, and then iterate, iterate, iterate. The similarities to how code is developed today are rich. This ‘sprint’-like methodology ensures that marketers can bring the rest of the organization with them on the journey.
Is this where the math comes in?
At some point, our brains aren’t built to understand the complexity of exactly how the machines develop their learning, but if we set up guardrails, we will continue to be the best guide for them to carry out their tasks. We have to think differently about the challenges we give them and focus on what we’re really good at—strategic thinking, creativity, emotional input, understanding the psychology of our customers. All those things which machines aren’t good at (yet).
What kinds of guardrails are you talking about?
In order for humans and machines to work together effectively, AI has to have transparency and explainability built in. We have to understand and account for bias (which isn’t as hard as you might think), to have insight into why the machines are doing what they’re doing and to guide them. While we want to let go of the reins a little bit and have the machines do what they’re great at in terms of automation, transparency into decision making to prevent unintentional bias against consumers for instance, is of course crucial. We need to build those into how we think about automating all parts of our organizations, and marketing is no exception.
We still haven’t talked about the cost…
Luckily, we are in the perfect time in history to realize the incredible benefits of AI. Firstly the cost of technology and processing continues to decrease at a rapid rate. Secondly, the large technology organizations and groups that once held progresses in AI and machine learning behind their walled gardens, have now increasingly open-sourced their work—meaning that AI is now democratized and simpler to access.
At Hux, one of our goals is to build on this movement, making AI capabilities more accessible and effective across organizations. While orchestrating decisioning across every channel might sound complex, with some of the assets we’ve built and the alliances we’ve created, we’re able to help make the complicated simple, allowing the entire organization to focus on the strategy of creating growth. In this way we can help our clients enjoy hyper-growth, by elevating the human experience at every interaction—an ability deeply underpinned by humans working more closely and effectively with machines.
Learn more about AI in marketing with Alex’s quick take.