Marketers, Stop Spinning AI
Is artificial intelligence the greatest thing to ever happen to marketing, or the worst thing to happen to humanity? Digital marketers seem to overwhelmingly believe the former is true, and many – without even fully understanding what it means, are already creating lines in the budget for it. In fact, a recent survey found that more than half of marketers plan to adopt AI over the next two years.
So, in the Context of Marketing, What Is AI?
Artificial intelligence is exactly what it sounds like: it’s the science of granting power to computers to perform actions that typically require human intelligence. It’s already in use in the digital marketing and advertising space – Google uses AI to interpret search queries, and social media monitoring tools use AI to understand the sentiment of conversations that occur online. Predictive analytics require AI, too.
However, although most marketers use AI in some capacity every day, the term does come with some unfortunate baggage. Data teams are reluctant to use the term AI, since they feel it evokes the fantastical: a future in which machines replace human consciousness. They prefer to define their work as 'machine learning.' Machine learning is technically a subset of AI, but we like to think of it as computational neural networks that help us deliver relevant advertising that converts for our clients.
Intel’s head of machine learning offers a great description of the difference between machine learning and AI – which are similar, but not quite the same: 'AI is basically the intelligence – how we make machines intelligent, while machine learning is the implementation of the compute methods that support it. The way I think of it is: AI is the science, and machine learning is the algorithms that make the machines smarter. So the enabler for AI is machine learning.'
While AI is having a day in the sun right now, interest has proven to be cyclical in nature. There was an 'AI winter' in both the seventies and eighties where a lack of solid outcomes produced a period of funding cutbacks and ensuing disinterest in continued research. Machine learning, on the other hand, is proving to have real application value and will continue to flourish.
AI in Digital Advertising – and What’s Gone Off the Rails
There are key areas in which ad tech companies tend to focus. Workflow and operational automation comprise one. In this area, the concentration is on enhancing the capability to traffic thousands of campaigns with minimal human intervention. Combating fraud is another area where AI comes into play. By providing data on which traffic converts for advertisers, data teams are able to apply machine learning algorithms on what inventory to buy.
AI is a top trend currently, and as is typical, it’s following the money. That unfortunately results in a somewhat baffling landscape for marketers to navigate. There’s a lot of confusion because many of AI’s capabilities in the digital sphere are still hypothetical and not 'ready for prime time.' While some, as previously stated, are already in use (AI is being employed in programmatic targeting), many capabilities are discussed and even marketed as if they are, but we’re not likely to see them in play for a few years yet. Marketers need to understand the capabilities machine learning provides, and what's actually available today. How exactly is data being applied to the optimization of digital campaigns? When exploring AI-driven technologies, marketers need to ask a lot of questions and make sure they fully comprehend the answers.
What’s Ahead for AI in Ad Tech
There are some challenges with respect to advancing AI in the industry. Staffing qualified people who understand the field is a challenge. Also, overestimating the capabilities will remain a concern. Everyone seems to think they’re going to revolutionize the industry, but perhaps we should all just focus on putting data to better and more sophisticated uses, thus driving better results for advertisers.
AI, and machine learning specifically, can have incredible results when applied to data sets – to effectively purchase inventory, dramatically reduce fraud and so much more. It offers massive value, and will continue to provide optimization and cost savings for the years to come.