MotiveMetrics' Chief Science Officer, Kyle Thomas, joined the AI4 panel on AI Driven Customer Acquisition, Aug 8, 2023 at MGM in Las Vegas. The full panel discussion is available HERE, but our favorite highlights are shown below.
Kyle opened the panel by sharing how the company started during his PhD research in Experimental Psychology at Harvard University. Initially, MotiveMetrics analyzed text from Social Media to develop customer profiles, understand consumer motivations and inform marketing communications. Today, in Paid Search advertising MotiveMetrics' focus is to conquer the "last mile problem" presented by Generative AI systems.
AI in Action: Kyle provides an example of leveraging AI to solve the problem of customizing marketing messaging to unique consumer market segments at scale. For a large review site aiming to attract more business customers AI helped identify hundreds of thousands of market segments and created customized ad copy that resonated with each audience to drive 65% more customers year over year. The improved spend efficiencies from AI-driven intent alignment yielded over $500K in customer acquisition cost savings over Q1-Q2 2023.
The Last Mile: The "Last Mile Problem" in supply chain management describes the logistical challenges at the last phase of transportation getting people and packages from hubs to their final destinations. The Last Mile Problem Is a good analogy for the critical short-fall of AI systems today. “It was easy when you got the trucks on the freeway, but how do you get them to navigate through the city. That’s a much more difficult problem. In fact, that’s a whole different set of problems. You can use the AI to get you 90 to 95 percent of where you want to go, but we intentionally build our systems with a human in the loop to address the Last Mile challenges.” When applied to creating marketing messaging even the most highly trained Generative AI models are often unpredictable, with wide ranging issues from managing brand guidelines to AI hallucinations.
It's Like 1,000 Interns: The future of Generative AI is more personalization at scale and the key is humans plus AI working together. Working with Generative AI systems on a daily basis it is obvious that not only are there many problems today that have to be solved, but individual AI models are better at certain use cases than others. As a result, the customization of AI systems is where the greatest need and opportunities are today. Kyle concludes the panel with a great analogy, "AI is not a genius. It's like having a thousand interns. So it's good, but it's not going to run your company for you. That's how I think about this stuff out of the box, and then it is up to you to customize it to get the most out of those 1,000 interns to get some actual business results".