Assessing the Intelligence of AI Marketing Tech

In 2011, IBM Watson had just beat champion Ken Jennings on Jeopardy! and Paul Roetzer decided to apply that same technology to marketing.

In an episode of The Intelligent Business Show, Roetzer, founder and CEO of PR 20/20 and the Marketing Artificial Intelligence Institute, chatted with show host Matthew Grant about his years-long journey of discovery to define and understand artificial intelligence (AI) and to establish the current and future potential of AI for marketing.

By 2016, marketing automation tools had hit the market and were being deployed with gusto, but they weren’t truly smart, and frankly, were barely automated. Stimulated and tickled by the irony that marketing automation companies sold mostly manual solutions, Roetzer set out to discover truly intelligent marketing technology that could learn and get better on its own.

Paul Roetzer

He wanted to make AI applications for marketers and their operations approachable and actionable.

“We’re not data scientists, we’re not programmers, we’re not AI engineers — we’re marketers,” he said. Making sense of AI would enable him to articulate what “AI” really means, does, and how it can be applied.

But pulling the industry along for this journey and helping the marketing space to understand it was —and still is — no walk in the park.

It took Roetzer about one year of writing and speaking about AI to actually comprehend it, and his journey to understanding is not unique.

“You could ask 10 experts and they don’t know! Everyone is literally defining it differently, so once you become confident enough to say, ‘Well, I understand it in my terms,’ then you’re good,” he said.

Sometimes ‘Smartech’ is Just Martech with Good Marketing

Grant pointed to the sheer volume of marketing solution vendor websites touting their AI, machine learning, and predictive analytics-fueled martech that promises to enable clients to identify intent from intent signals and/or use those signals to personalize messaging.

“If the people who are selling this don’t understand what those things mean, is it just hype? Or BS?” he asked.

Roetzer wants to chalk it up to industry-wide earnest ignorance of what AI and “intelligent” technology truly is. When marketers and sales reps at these allegedly AI-powered companies must market, sell, and develop messaging for tools they fundamentally do not understand, they write and sell to the best of their abilities.

Even at massive platform companies, he said, there may be teams of 300 engineers, but they’ve only got 3 people who actually know what machine learning is, how to use it, and how to identify and solve machine learning problems. Hence, platforms are built by folks who don’t know what AI is and who continue to “improve” it with capabilities that they don’t know aren’t intelligent.

Boiled down, Roetzer said, machine learning is making predictions based on historical data about future outcomes; but the difference between machine learning and data science is that the machine learning algorithm learns, gets smarter, and continuously improves its predictions and recommendations.

So, Does It Get Smarter on Its Own?

If a vendor’s software does not get smarter on its own, if its recommendations “don’t get better in real time based on new data, then it’s not machine learning,” Roetzer said.

An unfortunate side effect of this proliferation of not-truly-intelligent marketing technology — smartech, is i may, and i will— is that marketers don’t know when they’re buying quasi-intelligent martech. Therefore, they don’t know to push back and demand software that is truly intelligent.

“There should be demand for truly intelligent software, but the average marketer wouldn’t know it if they were staring at it.”

His litmus test for supposedly intelligent marketing tech is this: “Does it get smarter on its own?”

If it doesn’t, it’s not machine learning, and it’s not intelligent.

The Future of AI for Marketing

Attainable use cases for AI martech include content intelligence, budget allocation intelligence, and “things that require a lot of data and a lot of time,” such as advertising strategy and execution, Roetzer said.

He projects that in the next 3 – 5 years, there will be “ways to do things so much more intelligently, that as a marketer, you can’t fathom going back to setting up your own workflows.”

Grant wondered about “AI going the other way” — as in, “recipient-side AI that might make it harder to market to people.” He said, for example, how Microsoft Outlook divides his incoming email between “Focused” and “Other” folders, or how an email security management provider filters out marketing emails.

Roetzer reflected on personalization, privacy, and generational differences.

“The bet that the industry is making is that the convenience of having personalization is greater than the risk of your privacy being invaded.”

Some aspects of marketing will get harder the smarter AI gets, he said, but if the benefit and convenience of personalization win, he doesn’t think this will be a debate for younger generations, and suspects the ubiquity and evolution of AI will be a far greater asset than hindrance to marketers.