View all the CX Decoded podcasts.
The use case for Artificial Intelligence (AI) in the workplace is there. Deloitte’s Tech Trends 2021 found AI and machine learning technologies are helping financial services firm Morgan Stanley use decades of data to supplement human insight with accurate models for fraud detection and prevention, sales and marketing automation, and personalized wealth management, among others.
For marketing and customer experience, in particular, organizations are using AI and machine learning to improve internal business processes and workflow, automating repetitive tasks and to improve customer journeys and touchpoints, among other use cases.
The CMO Survey by Duke University reports a steady increase as far as the extent to which companies are reporting implementing AI or ML into their marketing toolkits. But it’s not exactly earth-shattering growth: on a scale of 1-7, one being not at all to seven being very likely, it’s at 4.1 as of Feb 2021 vs. a steady 3.5 over the last two years.
However, the majority of marketers know AI is very important or critical to their success this year, according to Paul Roetzer, founder and CEO of the Marketing AI Institute and PR 20/20. Roetzer said when asked how important marketing AI is to their success over the next 12 months, his research found the majority (52%) said it is very important (37%) or critically important (15%).
Roetzer caught up with co-hosts Rich Hein and Dom Nicastro on the latest CX Decoded Podcast to discuss the ways marketers are successfully implementing AI into their marketing campaigns and workflows.
Have a suggestion, comment or topic for a future episode? Drop us a line at [email protected].
Note: This transcript has been edited for space and clarity.
Dom Nicastro: Paul, we'd like to talk a little bit about what the Marketing Artificial Intelligence Institute does. Do you guys have a suite in the Salesforce Tower, or do you just do your work from your kitchen like the rest of us?
Paul Roetzer: I started the Institute as a blog, underneath my agency, PR 20/20. Some people may know me from that. We were HubSpot's first partner back in 2007. So we started working in the marketing automation space and the tech space pretty early on. I started the agency in 2005 and just sort of had a front-row seat to what was going on in the tech space. I wrote the Marketing Agency Blueprint in 2011, and that was the year IBM Watson won on Jeopardy.
And I just kind of took a look and I was like what is that? How does that tech work? Is that going to affect marketing at some point? How would we apply that predictive modeling and things like that?
And it just started me down this path of research and discovery. And so in 2016 we had been thinking about it for five years and it's like, well let's just start writing about it and see if any other marketers care about the topic. And that just sort of grew from there. We were up to about 5,000 subscribers in 2018, and I decided to split it off as its own business. We turned it into real media company and an event business in 2019. And that's kind of where we're at today.
Rich Hein: I like to think everybody in the audience is super technical. But clearly they're all not. Some of us are marketers who are copywriting. Some of them are more technical. But I think there's a lot of confusion around the main terms that we hear a lot, and I think they get wrongly used interchangeably. So I was hoping you could shed some light on AI versus deep learning vs. natural language processing and so on.
Paul Roetzer: It's always the best place to start, to level set. As you alluded to there are dozens of definitions, and depending on who you talk to, you're going to get a really technical definition or skew the other direction.
But after about a dozen years of trying to figure this out myself, I've come to have a favorite. And that is Demis Hassabis, who is the founder of DeepMind. If anybody watched the AlphaGo documentary on Netflix or seen it on YouTube, that is where they built the machine that won at the game of Go. They got acquired by Google years ago for $650 million.
So Dennis defines it as the science of making machines smart. That's how he defined Artificial Intelligence. So, what that means is machines, that computers we use, the software, can't do anything on their own that they're not programmed to do. If you think about your job as a marketer, you write all the rules, you figure out the emails to send, when to send them, what to say, what promo to offer, what price point to use. You are going through and literally writing all these rules that's trying to predict what's going to generate the outcome you desire.
And so that is the history of marketing. We're trying to do these things to get someone to take an action, and we don't really think about it as a bunch of all these predictions, but that's really what it is. And so AI is sort of this umbrella term for the tools and technologies that make machines smart.
Underneath that is the primary subset which is machine learning, and machine learning is literally what it sounds like. Instead of the human writing all the algorithms and telling the machine what to do, the machine learns as new data comes in. And it can actually write its own rules, and it can get smarter.
So as you'd alluded to up front with Waze and Google Maps and Spotify and Netflix and Amazon, what's happening is machine learning is behind the scenes, learning your preferences, your interests, learning the things that are going to get you to take the action they want you to take. So machine learning is the primary form of AI that is infused into a lot of consumer experiences and being built into a lot of software.
The subset of machine learning below that is deep learning, and that is actually trying to get the machine to, in a way, be human-like. To give it human-like abilities of language, of vision, of prediction. It can't do any of those things. It can't understand language. It can't generate language.
So AI is the umbrella term; the primary subset is machine learning; and the more advanced form of machine learning is deep learning. And it's basically trying to, as you alluded to, intelligently automate processes. So think of all the things you do that are repetitive and data-driven, it can do them for you. And then doing human-like things at scale like tagging photos or generating content. So at a high level that's kind of how I think about it.
Dom Nicastro: Paul, what have you seen in terms of how marketers are grasping the concepts of AI?
Paul Roetzer: If the litmus test is, you're a marketing manager or marketing director, and your CMO says, I keep hearing about AI, can you explain to me what it is and how we could be using it? If that's the litmus test, I would be shocked if the percentage of marketers who could answer those questions intelligently right now with no research is higher than 5%.
So, I truly don't think that the marketers, even the ones who are using it, even in the big enterprises who are applying it in certain ways, I don't know that they really understand what machine learning is. Is it different than deep learning? What is natural language generation?
There's probably like seven to 10 fundamental definitions or applications you really need to understand as a marketer to understand what's possible. You don't have to build machine learning algorithms. You don't have to be a deep learning expert. But to understand what's possible, to look at problems differently, I would be shocked if it's more than 5% of the industry. Shocked.
Dom Nicastro: There's a lot of ground to gain then. Let's try to boost that to 10% next year.
Paul Roetzer: That's the opportunity. If you're a marketer, you're like I don't really know that stuff. This is scaring me. That's the point, though. We're at the beginning. Everything we're talking about today, everything else we will cover on this podcast, it's just the beginning. Everyone's at the same point you're at if this is all new to you.
Rich Hein: It sounds like there should be low-lying fruit.
Paul Roetzer: What we always tell people is, there are two fundamental ways to think about getting started with AI. One is, as you alluded to earlier, just make a list of all the things you do every day, every week, every month, and then have a column that says how many hours you spend doing those things. And then have a column that kind of indicates how repetitive is it. Is this something that seems like I would be able to automate?
And you kind of go through and start picking these things off that you're spending all this time doing or that your team is spending all this time doing. And then literally just Google AI for ... whatever that thing is you want to do. That's a use-case level way to do it. You're just finding tasks or use cases that you spend a bunch of time doing that someone may have built a tool to intelligently automate.
The other path is look at problems you have in your organization. So if you think about customer experience and you start thinking about our churn rate is too high or we're not converting enough leads or like bigger problems. How have we tried to solve this before?
And you look at all those solutions matrix and you say, well, are there smarter ways to solve this? Is there tech out there we're not even thinking about that might make us way better at our job help us be more efficient and smarter? And those are kind of the two fundamental ways to think about piloting and adopting AI.