Artificial Intelligence: Are We There Yet?

James Regan, CMO, MRP

We all use and depend on our good friends Siri and Alexa, but the reality for B2B marketers is that technology powered by artificial intelligence to make our jobs easier isn’t yet as prevalent as we’d like.

There seems to be some confusion in the marketplace about what qualifies as true artificial intelligence and what current practices are better defined as machine learning.

While artificial intelligence holds great promise for B2B marketing, we’re not quite there yet. Machine learning, however, is here, and is already delivering tangible ROI for marketers.

What is Machine Learning?

While the terms artificial intelligence (AI) and machine learning are often used interchangeably, true AI is defined by Forrester as the ability for a machine to sense, learn, think, interact and take actions like a human. Machine learning is more accurately defined as a subset or pragmatic building block of AI. It means the use of algorithms that learn and improve through automatic feedback.

Machine learning will likely be an important part of true AI, but it is also a useful tool on its own. Machine learning enables marketing teams to make sense of massive amounts of data from a multitude of sources, filter that data in real time with an algorithm and then use those datasets to inform customer engagement tactics – from display advertisements to direct mail to inside sales calls and beyond. However, outreach is still based on a massive set of assumptions solely based on a marketer’s experience or intuition, creating less certainty and more margin for error. Machine learning understands, absorbs and learns from how customers engage with and react to the marketer’s strategic plan, thus improving future predictions. Each round of marketing outreach becomes stronger than the last as the machine makes smarter and smarter predictions based on feedback it ingests each time customers engage – or refrain from engaging – with marketing touchpoints delivered to them. Right now, though, machine learning cannot replace a marketer’s experience and understanding of human behavior. When it gets to that level of sophistication, reducing the amount of decision-making on the marketer, we can call it AI.

Although the kind of true AI for marketers that can design a media plan, or do unassisted copywriting may still be in the future, machine learning is a powerful tool for marketers right now. It can build predictive models and enhance marketers’ ability to engage prospective customers and nurture current customer relationships.

Putting it Into Practice

Within marketing technology, some machine learning algorithms and processes are more common than others. Most marketers are likely familiar with regression analysis, the ability to classify data to determine which parameters are most likely to predict or optimize for some goal. The more metrics available for analysis, the more “confident” the algorithm will be, but a human being is still required to quantify and qualify each of those inputs, and to interpret and define the tasks and actions that will follow.

There are many machine learning algorithms. “Deep learning” is a term we are hearing a lot lately. It simply means connecting those output tasks to new machine learning inputs, and while it can produce strikingly powerful predictions with less understanding of the inputs than other methods, it still requires a human marketer.

For us, this means trying different tactics – display advertisements, email campaigns, direct mail campaigns, etc. –and seeking solid metrics for the results. A predictive platform can track changes in behavior and provide data that will allow the marketer to understand how best to reach particular customers or prospects.

The feedback loop offered by machine learning allows marketers to continuously improve their outreach, delivering smarter, more relevant marketing tactics that reach prospective customers wherever they are in the buyer journey.

Another direction for machine learning algorithms allows marketers to segment audiences by understanding what they react to, whether that be a visual or linguistic cue as specific as a prospect’s preference for the color blue or a text treatment versus an image, allowing you to show each one a separate advertisement.  Machine learning algorithms can potentially identify these kinds of creative parameters automatically.

Looking Ahead

At this point, B2B marketers still need to invest time making the initial recommendation for which content should be served to which targets. In the not-so-distant future, artificial intelligence will remove the need for that preliminary guesswork. When machines develop the intuition and assessment abilities of humans and can determine and automate the best engagement tools, marketers will then be using true artificial intelligence. True AI for marketers, once it becomes available, can do much of the planning and speculation required before launching an outreach plan, allowing marketers to deliver better, smarter tactics right off the bat.

For now, we can only imagine the millions of other implications of AI on B2B marketing and how they will impact our roles. As machine learning and AI technology continues to advance, marketers will be able to grow and amplify their influence.

[author] About the Author:  James Regan's decades of experience across sales and marketing in both blue-chip corporations and startup environments led him to found MRP with Kevin Cunningham in 2002. As MRP's CMO, Regan keeps MRP on the cutting edge of marketing services and works on the architecture and execution of MRP's internal marketing efforts. He also oversees the management of MRP's growing team of internal marketing managers, marketing coordinators and account executives, coaching, training and constantly challenging the team to excel. [/author]