Everyone is predicting the rocket-launch trajectory of artificial intelligence in the 2018 marketing universe. In December, the ANA voted “Artificial Intelligence (AI)” as their marketing word of the year, topping previous winners “transparency” and “content marketing.” Those who cast their votes for AI suggested that it isn’t just a current fad, but rather a foundational capability and way of engaging with customers that will change our daily lives. While true understanding of AI is still not yet mainstream, it is gaining enough traction that marketers know they need to start paying attention to it.
In fact, there are already a number of examples of how AI is changing the world of marketing. Amazon has long been using algorithms to make better product recommendations to their customers, based both on individual customer preferences as well as general preferences across like groups of people. Apple’s Siri and Amazon’s Alexa use a form of AI to respond to customer requests and function as a helpful digital assistant. IBM’s Watson and Adgorithm’s Albert are products that directly offer an AI solution to help improve marketing ROI. In one case study, Harley Davidson reported that their use of AI improved lead generation by 2,930 percent. These stories alone demonstrate that artificial intelligence is in it for the long haul.
However, while there are a number of great success stories about improved ROI, there aren’t as many stories that talk about the people who are going to be necessary to ensure that AI continues to be an effective marketing tool. While there is a certain ‘hands-off’ and constant learning reality to the concept of artificial intelligence, this doesn’t mean that you won’t need people involved along the way. From setup to model-building to interpretation of results, you will still need humans involved in your AI efforts.
Let’s first consider that artificial intelligence is implemented by building models or algorithms — which implies both a need for data to serve as inputs to those models and infrastructure to host them. This means a consolidated customer data warehouse (CDW), marketing technology to capture customer info, a smart analytics engine, and connected systems that can leverage all of this amazing insight easily across multiple channels. In most cases, this integrated data environment does not exist. Someone has to do the dirty work of setting all this up on the front end — a smart marketing technologist, or a systems integrator, or an AI expert (or probably all three). Good AI systems will ultimately be a reflection of how well they are set up in the first place, not just the horsepower of the underlying algorithms.
Okay, now that you’ve got everything up and running, it’s smooth and hands-free sailing from here, right? Sure — kind of. This is definitely the part where AI does the heavy lifting. It delivers messages. It captures responses. And it learns from those responses. As it ingests information, it becomes smarter, improving relevance and timeliness along the way. As long as you have a healthy respect for AI’s strengths (e.g., programmatic ad buying) and limitations (e.g., personalization at the individual level), you can generally keep your hands out of the system and watch it do its thing.
But that is the trick— for many important business applications, someone has to watch it do its thing. You still need a smart data scientist to compare what is happening to what you expected to happen, and to raise a flag when there is a variance worth paying attention to. Growth and improved ROI must be compared to what was happening before, or what would have happened had AI not been in place. Judgments must be made on how AI is performing, and those performance results are meaningless without some kind of baseline to measure against. These judgments require human participation.
There also needs to be someone responsible for noticing when something doesn’t seem to be working right. Not everything that performs flawlessly up front will continue to do so indefinitely. We’ve all been the victim of some automated campaign that lost its data feed, resulting in emails to <#FIRSTNAME LASTNAME#> or contains links that go nowhere. Some sort of advanced warning system should be built to make sure that when something breaks, you know about it right away. Most importantly, someone needs to be responsible for doing something about it when those alerts are received.
Finally, one more person is required to ensure that AI stays intelligent. While AI is constantly learning based on what has happened, it doesn’t yet have a way to adapt to changes that companies are anticipating making. Business strategies evolve, customer targets shift, and product mixes change over time. These things won’t always impact AI, but there is a chance they might and someone needs to make sure that any necessary adaptations are made. For example, if you have programmed AI to pay attention to conversion-based metrics but later shift to an engagement-focused business strategy, models and algorithms need to be updated.
The future of artificial intelligence looks bright, and we can expect some interesting changes as it becomes more relevant. However, although AI might be the marketing word of the year, it will be the people behind that artificial intelligence that keep it relevant for years to come.