Opinion: How the rise of robots can boost the human creative process
Moravec’s Paradox states that while machines are great at doing some things that humans find complex, they are bad at doing other things we find simple. If that’s the case, then we should really be asking ourselves: “What is the complexity in our business we can focus these machines on to make them optimal creative partners?”
In the same way that the engineering progress of robotics faces bottlenecks, so does commercial creativity. Our bottlenecks are defined by the “iron triangle” that you learn about in management school: higher speed, better quality and lower cost. We are taught that you can only have two of these, at the expense of the third. Our clients expect all three. Machine-learning AI and automation, partnered with strong creativity, now makes this possible.
3 ways to apply machine-learning AI to commercial creativity
Data-driven insights: Great creativity comes from the constraints of a tight and accurate brief, based on observed consumer behavior—not on how people say they feel or what they claim they intend to do. (After all, if you want a consumer to lie, just ask them a question.) Machine learning helps refine data sets of extraordinary scale to reveal signals amongst the noise, moments that matter in consumer journeys, and help us understand the actions required to change the way people feel and act. The application of machine-learning AI to data generates invaluable insights, which facilitates more-effective creative.
Optimization and efficiency: The second way is the ability to fail faster, learn faster and fix faster, in order to get your creative expression exactly right. Long gone is the relevance of IPSOS-like testing of 30-second spots against a set of norms and zero context. Today, we can make, run, iterate and learn. In a recent campaign for a luxury car client, we used a machine-learning AI platform that can make subtle adaptations to creative based on audience data signals, down to the individual level. We set the creative parameters for the execution—images, sounds, claims and the overall narrative—then let the machine optimize it. The platform developed 438 permutations of the video, three of which significantly outperformed the others. The result was a 42 percent higher view-through rate and 17 percent cost reduction of the next-best action. The nuances in the creative wouldn’t necessarily have been apparent to a human, but by embracing AI, the work became more effective.
Outcomes evaluation: The first two examples are tied to traditional media metrics, which are important, but not the ultimate goal. To take it a step further, robots are also able to determine and measure the true attribution of creativity on clients’ business outcomes, and on the resulting behaviors of individual consumers. With this attribution it is possible to demonstrate how creative work solved those “jobs to be done” in the consumer journey from the brief. And, because it can be measured and its marketing impact is assured, it allows more clients to shift to an outcomes-based pricing model, leaving the inputs-based approach where it belongs—in the last decade.
These are just three of the many exciting possibilities for applying machine-learning AI to commercial creativity.
By embracing the inhuman to handle those complex tasks that distract and slow down the process, creatives can liberate themselves to become more human in pursuit of what was arguably the first, and will be the last, bastion of humanity—our ability to create.