Are you confused about what’s real and what’s hype when it comes to artificial intelligence (AI) and marketing? Well, join the club.
In 2017, Gartner revealed that more than 1,000 vendors claim to either sell AI or add it to their products. In the 2018 Marketing Technology Landscape Supergraphic, Scott Brinker eliminated the artificial intelligence category because AI is embedded in so many products across all martech categories. But is it really?
Gartner also said many tech vendors were "AI washing" by applying the AI label too indiscriminately. A survey by EverString and Heinz Marketing concurred: “AI is one of the most hyped terms in marketing technology with a variety of ideas and perspectives coming from vendors, thought leaders, and analysts.”
In short, it may be a stretch and disingenuous to say your company is an AI brand just because you make use of certain technologies.
What Exactly Is AI?
Even the definition of AI is open to interpretation. Frankly, I struggled to understand what qualifies as AI while researching this article.
Some say AI is “any technology that enables a system to demonstrate human-like intelligence.” Some take it a bit further, describing AI as “the presence of using algorithms to process data, and suggest an action.”
Others take it to another level, such as Alexandre Gonfalonieri in his recent Medium article: “AI analyzes data, makes assumptions, learns and provides predictions at a scale and depth of detail impossible to replicate by humans.”
Adding to the confusion are the numerous technologies that fall under AI. First, we have machine learning (ML), which is sometimes defined as "the ultimate in AI." ML calls upon mathematical models trained on data to make decisions and makes better decisions over time with more data. Some say AI can’t exist without machine learning, but machine learning can exist without AI.
Then there’s predictive analytics, an AI-based technology that makes predictions based on past data and patterns to predict what may happen in the future, but typically on a limited scale compared to AI.
There’s also natural language processing (NLP), which analyzes human speech and text (think conversational chatbots and virtual assistants).
Some vendors claim to offer AI-powered tools when, in actuality, they’re selling advanced analytics. Depending on whose definition you go with, unlike advanced analytics, true AI gets more intelligent over time, learning and adapting to data without human intervention.
To underscore this mass confusion, the EverString/Heinz survey found that 32 percent of B2B marketers aren't confident in their knowledge of AI, and 54 percent are only somewhat confident in their knowledge of AI. Only 12.8 percent of respondents reported feeling very confident in their level of knowledge of AI in the context of marketing technology.
This bewilderment may help explain why just 18 percent of those surveyed by Demandbase use AI for either marketing or sales.
You Don’t Need A Shovel -- You Need To Dig A Hole
Marketers know better than to simply chase the next shiny object, but we do anyway. Vendors know how to sell us the dream.
While Gartner predicts AI will be pervasive by 2020, marketers may want to validate a vendor’s claim that their product is truly built on AI. In many cases, what is being called AI is merely technology calling upon predictive analytics or simple algorithms.
That said, what’s most important is first deciding why AI is needed. Just because vendors are hyping AI doesn’t mean marketers truly need it. The first order of business is confirming that the technology solves a business problem for you.
Next, it’s critical that marketers have enough data to make use of AI. Without it, the AI model won’t work correctly or effectively. In fact, it’s quite likely the technology you’re considering isn’t true AI if the tech vendor downplays the importance of data.
Keep in mind that it takes considerable time to acquire and clean that much data. Yet, 71 percent of B2B marketers cited old or outdated data as one of their top three issues in a recent survey by Demand Gen Report, and 62 percent don’t have the time or resources to implement an effective process to improve their data quality. Additionally, research commissioned by Dun & Bradstreet shows that only half of B2B firms surveyed believed their customer data was complete, accurate and trustworthy.
Moreover, true AI requires a human to train, test and tweak it. Someone needs to submit queries for the AI to analyze and needs to authorize AI to take actions. That requires expertise because AI can cause harm or discriminate when poor quality (or biased) data is used during training.
Before you purchase any AI solution, make sure you can integrate it with your existing martech stack. It’s not a worthwhile investment unless you can use it seamlessly with your existing solutions.
Perhaps the real question is whether every company is an AI company (or needs to be). Marketers don’t need AI embedded in every solution they use; they simply need solutions that solve their most pressing business needs -- whatever that technology looks like.