Paul Roetzer, founder and CEO of the Marketing AI Institute, said that marketing AI is the “science of making marketing smart.” Notice he didn’t say “automate things.” Marketing AI is not marketing automation, and therein lies an important point – a lot of marketers still don’t understand what AI brings to the table for marketing.
Roetzer’s Institute, in partnership with Drift, a provider of conversational AI solutions, recently published 2021 The State of Marketing AI. It’s a survey that looked at how marketing AI is in use today, the most common use cases, and where companies see themselves using AI in the next five years.
We’re only getting started
Yes, we’ve been having lots of conversations around AI (Artificial Intelligence) for years. But marketing is slow to catch up, and the report makes that clear. Fifty percent of respondents considered themselves beginners in terms of understanding AI terminology and capabilities. Another 37% said they were intermediate in their knowledge of marketing AI. It’s a similar story when evaluating AI-powered martech - 64% have medium to low confidence in their ability to assess and select AI technology.
For the study, 38% of respondents were C-suite, primarily the CEO or CMO, and most were B2B, with some doing both B2B and B2C marketing. In addition, the bulk of the respondents were involved with content marketing (69%), marketing analytics (60%), and email marketing (58%).
Most companies are just getting started with marketing AI; 65% categorize their stage of transformation as “researching.” These companies are building their awareness of the importance of AI and how it can transform their talent, tech, and strategy. However, they aren’t doing anything because they don’t yet comprehend what they really can do with it – all they know is that it exists in different forms.
The next biggest stage is “understanding.” This group of respondents is learning how AI works and exploring the various use cases and technology. There are three other stages of transformation: piloting use cases, humanizing by integrating AI and human capabilities and scaling for wider adoption.
Respondents were asked what the barriers to adoption of marketing AI were for them. Seventy percent said a lack of education and training, another 46% said awareness and resources, followed by talent, skillsets, and strategy. Technology infrastructure was number 7 on the list at 35%. These barriers make sense if most companies consider themselves in the research and understanding stages of AI transformation.
Something else that aligns here: 77% have less than one-quarter of all marketing tasks automated, yet 80% believe more than 26% of tasks will be automated within the next five years. There is an expectation that things will change - and by the looks of it, they will change fast. The technology is available today to make the shift to marketing AI, but each one supports different use cases.
If companies are in the fast lane of marketing AI adoption, what primary use cases will we see first?
Marketing AI use cases
The Marketing AI Institute categorizes marketing AI use cases into five categories and over 45 use cases:
The Institute provides an AI scoring system that any marketer can take to see what uses cases should be their primary focus based on their situation and goals. The questionnaire asks you to rate the importance of over 45 use cases where AI could positively impact their business from 1-5. Then, depending on the outcome of that ranking, it will identify the top use cases selected and a set of technologies that support those use cases.
You might not realize how many ways you can improve your marketing through AI, whether it’s use cases that help increase revenue or reduce costs. I was surprised to go through the list and see all the ways you can leverage AI technology. Just reading through the list explains why so many marketers are still a bit lost with marketing AI.
Here are the top five use cases respondents in the report selected:
- Recommend highly targeted content to users in real-time
- Adapt audience targeting using behavior and lookalike analytics
- Measure ROI by channel, campaign, and overall
- Discover insights into the top-performing content or campaign
- Create data-driven content
Of the top ten use cases, four were content-based, meaning marketing was looking for ways to improve or write content using AI technology. This makes sense considering many respondents were involved with content marketing.
Maybe some of the top 10 use cases are quick hits and visible examples of the value of AI. After all, leveraging AI for planning isn’t something you can easily show executives to win buy-in for new technology and strategies. But show them how AI can create good content or deliver personalized experiences at scale without requiring additional resources or budget, and chances are you get that buy-in pretty quickly.
If the whole idea of marketing AI confuses you, I would recommend reviewing the list of use cases in this report. Seeing examples of where AI can improve a marketing strategy or activity helps clarify what the AI will do. Once that’s clear, it’s easier to look at the technology that implements that particular AI.
There are plenty of surveys available on the use of AI in marketing. For example, HubSpot asked marketers about their use of AI in its State of Marketing 2021 report. That report said that marketers’ use of AI had increased 190% between 2018 and 2020, while the 2020 State of Marketing report from Salesforce said use of AI in marketing had increased 186% from 2018. These two reports paint a different picture from above - that more marketers are actively using AI to improve their marketing strategy and activities. But the overall point is the same - marketing AI is necessary for success.
I’m not sure it matters how many marketers are actually using AI today. The more important point is that AI has the potential to help marketers accelerate revenue by finding the best-fit customers and delivering the right content and experiences. And it can help reduce costs by taking on some of the routine marketing activities.
We definitely need more education and understanding of what AI can do and what technologies are available today to help.