Artificial intelligence (AI) has been around for a long time. But over the past few months, it has gathered pace and taken the marketing industry by storm, providing marketers with unprecedented opportunities while opening more doors on how to distill its power to full capacity. Although we are now seeing iterations of the use of AI, some questions remain around the technology.
In the world of marketing, MARKETTECH APAC We wanted to discover how brands can use AI to succeed and, in doing so, what challenges they can prepare for, as well as the opportunities that can be achieved through AI innovation. For this, we turn to Konrad Feldman, CEO and co-founder of Quantcast– advertising technology specializing in real-time AI-driven advertising, audience insights and measurement.
In the last Expert up close interview, we sat down with Feldman to get his expert take on the emergence of AI and how marketers can use it as a powerful tool to ensure excellence in their marketing initiatives.
First, according to Feldman, AI innovation has exceeded its limits, it is not just “one thing” but a set of approaches. What makes it a real marker of growth in the digital age is the fact that it can interact with what has been democratized for everyone; information on the open Internet.
Right off the bat, productivity is the biggest benefit of AI, and for marketers, that means being able to have more space to do the job they’re meant to do better.
“There are going to be all kinds of companies launched and tools that build on these models (in broad language) that will help us improve our productivity and help us get our work done faster, freeing up specialists from marketing to be more creative and effective. more innovative. I think everyone should try these things to get an idea,” he said.
Konrad also added that the more we can configure these tools to understand data patterns so we can predict the right audiences and optimize campaigns autonomously, it frees people up to do things they still do well better.
“One of the most complex use cases of AI for marketers is being able to refine which audience segments are best to reach for an advertising campaign,” Feldman said.
“So one of the main benefits (of AI) is to help decide which consumer group would be best to reach for an advertising campaign. In any market there is significant potential (audience), but the reality is that very few marketers want to reach the entire audience.
Many advertisers have an increasing amount of information about their customers. Feldman gave the example of an airline that offers a route to San Francisco and is promoting tickets to that destination. She might have many guesses that someone might become a customer. But even with a large number of people, it would only be possible to identify a subset of these motivations.
And that’s where programmatic advertising comes in and leverages machine learning, which can more systematically assess the characteristics of customers who might be interested in a brand’s offering.
Ultimately, Feldman said there are different types of AI for different types of problems, but like any other technology, one shouldn’t start by just using the technology.
“You should start by solving a specific problem,” he said.
Overall, machine learning and AI algorithms work as an optimization process in which they attempt to minimize certain errors and maximize a certain value. We must therefore be able to incentivize the way the algorithm learns.
“Be very clear about the problem you want to solve and how you will measure success. I think that’s an important aspect: having a clear understanding of what success looks like.
Another important thing is the willingness to experiment.
“Be willing to experiment, recognize that not everything you try with new technology will always work right away. »
He explained that if something doesn’t work and you can understand why, that’s how you learn in advance.
He then concluded: “So it’s powerful to have a model where you’re able to experiment and learn quickly. And that’s the last thing I would say: experiment. These new products available based on (breakthroughs) and availability of data are such that we see some really interesting emergent properties from these models. And they are available and accessible. And everyone can use them. Try them!
Watch Feldman’s full interview here.
