Does a landscape for artificial intelligence (AI) marketing vendors exist? You’d think so, with all the hype for what AI and machine learning can do for marketers. And, after all, half of companies are using some form of AI in marketing initiatives, according to a Deloitte report last month. So where is that formal grouping of such vendors? The Gartner Magic Quadrant? The Forrester Wave? Even Scott Brinker’s Martech Supergraphic has no category for AI vendors.
Of course, there are hundreds, probably thousands, of vendors claiming they can power marketing processes, campaigns and content with some form of AI or machine learning. "Virtually every marketing technology vendor has an AI story," Gartner researchers Andrew Frank, Mike McGuire and Jason McNellis found in their Cool Vendors in AI for Marketing report published last October.
But is AI in marketing at the capability stage, or do vendors provide all-in-one AI marketing suites? You don't see vendors branding themselves as "drag-and-drops," right? They're a Web Content Management (WCM) vendor that has drag-and-drop. It's one of many questions marketers need to ask in their AI marketing vendor selection process.
It's Going to Take Some Work
So, what can marketers do when they're in the market for "smarter" technology for their martech stacks beyond Google Search? They need to work, and it's a lot. For marketers on the hunt for “smarter” technology, nothing will come easy in terms of selection. If you’re a marketer thinking of investing in AI capabilities, experts told CMSWire you’ll need a deep, thoughtful selection process that includes knowing how smart your vendor’s technology actually is, who’s on their team for engineers, what specific uses case you want to improve and knowing the impacts on your marketing organization in the near- and long-term.
“You really have to be able to ask different questions to these vendors,” said Paul Roetzer, founder and CEO of PR 20/20 and creator of the Marketing Artificial Intelligence Institute. “At the end of the day, is it smarter technology than what you have now? You don't go buy AI technology. You're not trying to find an AI solution. You're trying to find a smarter solution to a problem you have or an activity you're trying to execute.”
Related Article: Cutting Through AI Marketing Hype: It's About Machine Learning
Where Marketers Find AI Valuable
There is no denying marketers have already invested in AI capabilities for their marketing stacks. Gartner’s 2018 Marketing Technology Survey found 41% of marketers use AI or machine learning to support predictive analytics. That was the number one answer ahead of:
- Marketing campaign decisioning (34%)
- Text analytics (31%)
- Personalization (30%)
- Prescriptive analytics (27%)
- Diagnostic analytics (27%)
- Conversational user interfaces (26%)
- Object detection/recognition (23%)
- Speech recognition (23%)
- Machine translation/localization (23%)
- Paid media optimization (20%)
- Logo detection/recognition (19%)
- Facial detection/recognition (19%)
- Emotion detection/recognition (16%)
And, according to the IDC, the AI software platforms market experienced steady growth in 2018, growing 26.6% to $2.6 billion. “AI thrives in conditions where there is an abundance of cause-and-effect data, a large number of possible actions, and little time to analyze complex decisions. Marketers routinely struggle with these kinds of conditions, so marketing is fertile ground for AI solutions,” Gartner found in their "Cool Vendors in AI for Marketing Report." Gartner profiled what it found as “new and innovative vendors” in that report: conDati, Course5 Intelligence and YouScan, by no means, in Gartner’s own words, an exhaustive list. But it does demonstrate there are vendors focusing on AI in marketing.
Recognize the Iterative Experience
Yet still, despite the “cool” technology emerging, marketers have a long way to go — and a lot of work to do — in truly understanding what AI and machine learning can do for their marketing processes and campaigns, according to McGuire. “Don't get totally disillusioned with this,” McGuire cautioned marketers about investing in AI capabilities, “This is going to be an ongoing, iterative experience. You’re going to find some applications of AI, but it’s a matter of maintaining balance, if you will. If you're a marketing technology executive, keep your eye on the long term, don't get caught up in the negative stories, but take them in context… We have to figure out as marketers and as organized societies how we want to deal with this technology. We're still kind learning as we go.”
Related Article: Establishing AI Ethics in Marketing
AI in Marketing Won't Change Everything
The first thing in integrating some form of smart software into your martech stack is to recognize you can’t just build around AI to “change everything,” according to McGuire. “It needs to learn before it can work with us across all of our marketing disciplines,” McGuire said. “But as far as an investment, if you've got an existing martech stack, adding some of those AI capabilities can be really particularly useful, especially if we can start reducing very manual processes we've done in the past.”
However, he added, marketers need to understand that just because your new smart technology builds around a core use case, it may not be ready to integrate with an execution system that sends messages, for instance. “That’s important for the buyer to understand: if we’re starting with it as the core and we're building some of our other marketing requirements on top of that, that could take a while. So it's about trying to understand what it is you're really trying to accomplish with it, because … it’s not going to fix everything right now.”
How Will the Tech Impact People, Processes?
That comes back to having a deep understanding of your processes, existing skill sets within the marketing organization and beyond and resources needed to execute integrating AI into your marketing stacks, according to McGuire. If you don’t try to understand how AI in marketing will change the nature of the work and the processes in your organization, you're going to end up imbalanced, McGuire said. “You may have too few or too many or wrong types of folks or you haven't trained your existing staff enough to understand."
The good news? Companies are recognizing the need for this kind of talent. Hiring growth for AI and machine learning roles have grown 74% annually in the past four years and encompasses a few different titles within the space that all have a very specific set of skills despite being spread across industries, including artificial intelligence and machine learning engineer, according to the LinkedIn's 2020 Emerging Jobs report. Skills in those roles include machine learning, deep learning, TensorFlow, Python and natural language processing. Do your marketers have access to engineers with those skill sets, or possess those themselves? “We have to be able to look at the implications on the organization and the processes," McGuire said. "… How is that going to change my marketing organization? Is this going to mean less emphasis of me having to use agencies for a lot of my creative? What am I going to need in terms of my internal staffing? Do I just need to hire a bunch of data scientists? Probably not. Looking at our campaign directors, how are their roles going to change based on our utilization of this maturing technology?”
Related Article: 3 Misconceptions About AI in Marketing
Can Your AI Marketing Vendor Explain AI?
Those aren’t the only questions you should be asking. Vendors need to be vetted thoroughly. Likely, this is advice you’ve heard before in selecting marketing technology. But with those who promise marketing outcomes through AI or machine learning, too often there’s a “disconnect” between what vendors promise and their ability to explain the type of AI they offer, according to PR 20/20’s Roetzer. Remember that MMC report that found 40 percent of European startups billed as AI companies don’t actually use AI "in a way that is 'material' to their businesses"?
You've got to know what's behind your AI marketing vendor's technology “because all machine learning isn't equal,” Roetzer said. McGuire of Gartner suggested marketers be on the lookout for developments in "explainable AI," something that Google champions, among others. Roetzer cited the example of content intelligence vendors. “Almost all of them are going to use some variation in their messaging of AI: natural language processing, machine learning,” Roetzer told CMSWire. “But then when you talk to that vendor, very rarely can the salesperson actually explain what any of that means. They know they do content strategy. They know their tool enables you to figure your content strategy out. But they can't tell you what machine learning piece of it is. And they can't explain natural language processing and they don't really even know the definition of AI… Most of their sales reps and marketing people don't understand AI. And most marketers don't know how to simplify what they're trying to get out of the vendor.”
Specific Use Cases Always the Best Route
Therefore, if you’re a marketer and you do content strategy for a living, go to your potential vendor with very tangible use cases: I spend 25 hours a month, figuring out what topics for our writers to write about. And here's the process I go through right now. How does your software make it smarter? “It’s great if they're using natural image processing and machine learning and whatever they want to explain it as,” Roetzer said. But is the tech making your process “smarter,” in other words, do you save tons of money with it? And does it give you a greater chance of achieving your goals? “That's really the only reason you would buy AI technology no matter what kind of marketing you're doing,” Roetzer said. “It needs to save time and money by eliminating repetitive data-driven tasks. And it needs to improve your chance of achieving your goal.”
Grill Your AI Marketing Vendor
No one actually assesses how smart the AI tech is when examining a vendor, Roetzer added. The biggest problem in the future of buying AI marketing software, Roetzer said, is knowing which software offering is actually smarter. “As a marketer, do I know who is actually smarter and is going to save me the most time and money or give me the greatest chance to achieve my goals and answers?” he asked. “There is no way to do that right now. You have to as an individual marketer, learn all this stuff to be able to properly assess vendors.” What does the founding team look like? Who’s on the team? How many AI engineers do they have? “How many machine learning engineers do they actually have on staff because there's a lot of vaporware,” Roetzer said. “It’s like they're claiming to do this and then you dig in and you're like, you have two engineers and neither of them are machine learning engineers? That is how are you planning to do advanced AI? Who’s doing it? Because nobody on your team appears even be an AI person. And there's a lot of that but the vast majority of marketers would have no idea what questions to ask a vendor that's claiming to have AI.”