AI For Marketers: Practical Uses Across The Customer Journey

Tessa Burg: Hello, and welcome to another episode of Lead Generation brought to you by Tenlo Radio. Our guest today is Paul Roetzer. Paul is the CEO and founder of Marketing AI Institute and PR 2020. Paul, thank you so much for joining us. We’re excited to jump into this conversation today.

Paul Roetzer: Great to be here, looking forward to it.

Tessa Burg: So AI is a very large space, and we only have 25-ish minutes to talk about the impact it potentially can have on marketing in 2022.

Tessa Burg: So right at the top of this conversation, I’m gonna let everyone know that if you are interested in the content that you hear today, and especially if you’re looking for a way to accelerate generating leads to accelerate your pipeline, definitely check out MAICON, which is a Marketing AI Institute Conference, and it is August 3rd through the 5th in Cleveland Ohio, and to learn more about that and to register, you can visit marketingaiinstitute.com. So that’s Marketing AI Institute, all spelled exactly as they should be, no fancy spelling, but all in one word, no spaces, no dashes, Google it, register, I went last year, absolutely fantastic and worth it.

Tessa Burg: And Paul also just shared that there are free trainings that you can register for to get more people in your organization aware, and educated and engaged with the benefits of AI. So definitely check that out. Paul, do you have anything to add in regards to the conference or the trainings before we start?

Paul Roetzer: Gonna be in-person in Cleveland, that’s the hope and the plan, hopefully no more virtual. There might be a virtual component to it, but back together again in August of 2022, that is the grand plan.

Tessa Burg: Yes. I really hope that that happens.

Paul Roetzer: You and me both.

Tessa Burg: Yes. So Paul, before we talk about, you know, what are some of the tangible things we can do with AI in 2022, give us a little bit about your background, and what you’re most excited about this year.

Paul Roetzer: Yeah, I’m actually an agency, well, former owner of an agency, I sold my agency last fall. So I started PR 2020 in 2005, became HubSpot’s first partner in 2007, and then sort of rode that marketing automation CRM movement, inbound marketing, digital marketing, we’re going to call it for, you know, 15 years or whatever. So that was my background, was an agency.

Paul Roetzer: And in 2011, right after I finished my first book, “The Marketing Agency Blueprint” I started trying to understand what artificial intelligence was. It was just a curiosity, really I’d seen, IBM Watson went on jeopardy, and I wondered what that was, and if it could eventually be applied to marketing in any way, this kind of idea of making predictions and creating answers.

Paul Roetzer: And so I set out on this journey to try and discover, could I use this thing, artificial intelligence, whatever it was, again, not really comprehending it, to help me build better marketing strategies, to allocate budgets more intelligently. And so it took years to try and comprehend what exactly AI was because the only people writing or talking about it at that time were AI engineers and data scientists who, no offense, suck at explaining artificial intelligence to the average person like me.

Paul Roetzer: So I just basically tried to understand it at a very simple level. What is it? And so I think the key for me was just this realization that it’s just smarter technology. It’s just the ability for the software we use every day as marketers, whether it’s email, or CRM, or automation, or social media, or advertising, to have that software get better and better, and help you make smarter decisions, make better predictions about outcomes or human behavior.

Paul Roetzer: And so if you just start thinking about, it’s like well, I already buy software, I already buy all this technology, if that tech on its own could continuously improve and make me better, why wouldn’t I wanna do that?

Paul Roetzer: And so really, I mean, we can get into kind of what is it, and how does it actually work, but at its core, AI just makes predictions about outcomes, about behaviors and it helps you understand and generate language and do these other things that as marketers, we’re doing all the time, there’s just better tech out there that you could be using.

Paul Roetzer: So for what I’m mostly about 2022, besides being back in person for a conference is having more marketers realize that AI isn’t abstract or overwhelming. It’s actually quite simple to understand, and it’s quite simple to apply it to what you do every day. And so for me to see more marketers understanding and adopting AI is what I’m really intent on helping move forward in the industry this year.

Tessa Burg: That sounds like a simple goal, but I also know from talking to some of our clients, it is quite ambitious, because even though I feel like you did a fantastic job at the conference last year at making AI very accessible, there are still a lot of stakeholders in companies that you sort of need to bring along to start actually using and executing it. Have you seen any companies who just do a fantastic job or what would you say to a marketer who knows the value, who wants to start executing it, what are some tips you could give them to sort of start getting that buy-in from other people in their organization, on their team, or at their executive level?

Paul Roetzer: Well, from an education standpoint, you had mentioned at the beginning, we’ve just launched this intro to AI for marketers, and it’s a free online class. It’s like, our positioning is you’ll understand AI in 30 minutes or less, like, and then do whatever you want with it. Like either take no steps or like immerse yourself, go to a conference, do all everything else. But our whole plan is like, we have to get more people to just understand the basics, and to realize it’s not sci-fi, this isn’t something in five years from now you build this, it’s like you could go find a tool right now for 19 bucks a month to write your tweets for you, like something as simple as that, or do lead scoring, which I know we’ll talk about. Like, you can find tools to help you do the things you’re already doing every day.

Paul Roetzer: And so that’s what we always advise people when you’re trying to get started, just make a list of the activities like the tactics you’re doing every day as a marketer or a salesperson, a business leader, and then go through and say, “Okay, how valuable would it be to me “to intelligently automate this task?”

Paul Roetzer: And AI doesn’t go from like zero to full autonomy. That’s not what we’re trying to do. You’re not replacing humans. You’re just trying to take a process that maybe has a few steps to it. And you’re trying to automate pieces of that, that thing, you’re not automating entire jobs away.

Paul Roetzer: So an example would be like, if I’m gonna send an email newsletter, I have to segment the list, I have to pick the time it’s gonna send, I have to write the subject line, I have to write the copy, I have to pick the CTA, I have to do all of these different components. How many of those steps could AI do for me, or maybe even do better than me? And like subject line writing is a really good example. There’s tech that will write subject lines for you that will outperform human written subject lines every time. And it’s because it learns from every other email you’ve ever sent. And so it takes all this data in, and it actually predicts what subject line will work best. It can do the same thing with body copy and emails, and what CTA to use, what image to use in ads, and all these things.

Paul Roetzer: So if you just make a list of all the things you do, and then say, “Okay, these are the five that I spend like “80% of my time on, “let me go see if there’s AI tools to help me “with any of these five things.” That’s your starting point. And you might find five different techs that have been built to do those exact things. And you might be able to get a free demo of them or get a one month trial or whatever it is, and just experiment with it. You can’t understand this stuff until you actually just run an experiment or sit in on a demo.

Tessa Burg: I think that guidance is perfect, because it covers, you already offer marketers resources to start with that education. So if they go to marketingaiinstitute.com, they can leverage those resources to get buy-in and awareness of what AI actually is. And then experimentation, since a lot of things are, you know, web-based and cloud-based and cloud accessible, you don’t have to overinvest to do something simple very quickly to prove out that value. So those are two fantastic first steps.

Tessa Burg: So what we’re gonna do on the rest of the call is help all of the listeners start to build out that list using the sales pipeline as our guide. So we’re going to give you some applied AI examples for prospecting, lead qualification, demo and engagement, and closing, and Paul’s gonna share with us, where can you get started? Where can AI start to add value to your marketing and sales practices? So let’s get started, prospecting. How can AI help marketers increase visibility and find the right customers?

Paul Roetzer: So, I mean, just, and we didn’t actually prep for these. So I didn’t like create a list of all these things

Tessa Burg: Oh yeah.

Paul Roetzer: But off the top of my head, I’ll tell you one I use all the time, which is LinkedIn Sales Navigator. So Sales Navigator, when you’re prospecting for contacts, so I’ve actually been doing it, trying to find podcasts. So this isn’t, I’m not using it for like lead gen and qualification purposes at the moment. But what LinkedIn Sales Navigator does is it uses AI algorithms to look at the context you’ve already identified as being good fit, and it’ll recommend people like them to you. So that is an example of passively using artificial intelligence to find better prospects for your business. So that’s the first one that jumps into mind.

Paul Roetzer: Another would be, it’s very common to use like lookalike lists, so if you’re running ads, like let’s say you create a piece of content, an asset that people can download to generate leads with, you know, rather than just going to the couple thousand people maybe in your database, you might be able to take a list, upload it to Facebook or LinkedIn or Google or wherever, and have a lookalike list built of other people like the people that you know are already good qualified leads. So again, another very passive example where you don’t really think about using AI, you’re just trying to find, you know, contacts that look like the people you had, but the way they’re doing that is it’s making predictions.

Paul Roetzer: So the core of AI is something called machine learning, and machine learning, what it does is it takes data in and predicts outcomes and behaviors. So in this case, it’s trying to find patterns like, it looks at all the different points within these contacts, and then tries to predict that these people look like these people. So that action of making a prediction and a recommendation are fundamental to how machine learning works, and what it enables for marketers.

Tessa Burg: I feel like light bulbs started going off in a bunch of people’s heads. They probably didn’t realize that they were already using AI, you know, like quite a few people have heard of, you know, LinkedIn Sales Navigator and have probably done the lookalike tactic. And so right now that gives them a little jumpstart on starting to communicate that value, and call out like, “Hey, this is a great example “of how AI has helped us improve performance.” and start to take that next step.

Paul Roetzer: Yeah, one of the things we’ll tell people, a lot of times I’ll start my talks with this is like “You’ve used AI dozens of times today “and didn’t know it.” And so to make it more like approachable to people, it’s like, okay, if you listen to Spotify and listen to a song that recommended, if you watched a show on Netflix that was recommended to you, if you took Google maps, you know, you used a route Google maps recommended. If you talk to Siri or Alexa, like none of these things exist without artificial intelligence. And in many cases it’s just personalizing experiences and making recommendations to you. That’s not possible at the scale these companies do this stuff without AI embedded within the algorithms.

Tessa Burg: Yes, and I love it it’s getting better and better. My Netflix recommendations are finally starting to improve.

Paul Roetzer: Netflix invests a ton of money and talent in that. I mean, these brands like Stitch Fix, we wrote about Stitch Fix in our upcoming book. They hire astrophysicists like away from NASA to recommend clothing. Like they build recommender algorithms for personal style. So, I mean, there are brands that are racing forward with applying AI to things you wouldn’t even think about.

Tessa Burg: That’s amazing. And also explains why I’m struggling to cancel my Stitch Fix subscription. I was like, this is the last box, but then I get the next one and it’s perfect.

Paul Roetzer: It gets better every time, that’s the whole point.

Tessa Burg: Yes, gosh darn it. So let’s move on to qualification. And you mentioned scoring earlier, and I think this is an awesome place for AI to play a role, instead of people arbitrarily trying to determine how much value they should be placing on activities. Tell us a little bit about how AI can help with qualification and scoring of leads?

Paul Roetzer: Lead scoring is one of the early efforts that a lot of, especially MarTech companies made to apply AI, so again, like thinking of the thing that I was like, AI is actually relatively new in its application to marketing and sales.

Paul Roetzer: When I first started researching in 2011, no one in marketing was talking about it. Maybe Salesforce was starting to dabble in this stuff, and starting to make some early acquisitions, but generally speaking, most of the major MarTech companies, weren’t doing anything. And then in part it’s because much of what’s possible today with language and vision technology and predictions that are actually really good are merged in the last like three to five years with some leaps forward in AI technology. So the reality is a lot of lead scoring sucked, like a lot of the early efforts to do it weren’t very good.

Paul Roetzer: That being said, it’s an ideal use case for AI, because what it’s doing again, like let’s say you have 10,000 leads that you can go through and say, “These are really good quality leads that we’ve had “over the last 24 months.” Now for a human to look at those 10,000 leads and say, “What are the commonalities here, “like what are the things that make them good?” We’re gonna do the obvious things. We’re gonna look at first their intent signals. So like, well, did they click on links? Did they visit certain pages? Did they, you know, take actions with emails that were sent? It’s like, okay, that’s pretty obvious, and you can set some rules around that stuff. Then you might look at geography. You might look at size of company, number of employees, revenue, funding, like you’re just gonna go through these obvious checklists of 15 to 20 things, and the current approach, which is build some rules, like a rules-based lead scoring model that gives X scope points per action or per signal. Well, when you start trying to scale that up though, and you have thousands of web pages, or you’re sending hundreds of thousands of emails, or you doing all these things, it’s impossible for the human mind to actually assess this stuff in real time and at scale. And so that’s where AI comes in as this ability to constantly monitor thousands of potential signals and to be able to figure out what does all this mean? It finds patterns.

Paul Roetzer: So pattern recognition again, is like a core thing that machine learning makes possible, and pattern recognition at that large scale is what it’s trying to do with lead scoring. So like a company that comes to mind would be MadKudu. It’s not a cheap technology, but like that’s a tech that a lot of people I know use that really like their lead scoring.

Paul Roetzer: The thing I will tell you though, is like, if you’re a small business or even if you’re a big company that doesn’t get a ton of leads, AI-powered lead scoring probably isn’t gonna be much help. Like you almost need to have, I’m just going to pick somewhat arbitrary number, but let’s say 2,000 leads per month now that you can actually, ’cause you have to train the AI in a lot of cases, you have to go say, “Okay, we had these 2,000 leads, “these seven ended up buying from us, “and they were really good.” The AI needs to know the end game, like the outcome. It needs to know what a good lead looks like, so you have to kind of teach it that, but then over time, it can now scale up its learning, and it can keep evolving and getting better.

Paul Roetzer: So lead scoring with AI, likely you need a higher volume of leads and conversions for it to be useful, otherwise you can probably get away from just your intent-based rules. It’s like, don’t overthink it, but you need a lead scoring model of some sort, no matter what kind of company you are. But in some cases, a human rule-based model might still be the most viable solution. AI isn’t always the answer.

Tessa Burg: I’m glad you mentioned that, because even if you were a large company, someone listening is like, “Well, yeah, I mean, “we get 2,000, you know, contact us forms, “so this works for us.” You still need to start with a person laying out that strategy and process and identifying what those activities are. And in qualifying, who is a good lead in order to the algorithm accurately, ’cause if you don’t identify the right features, then you’re also aren’t going to, and to have a strategy and a process and the data in place, then you just have a tool.

Paul Roetzer: Right. And the whole thing with AI is a lot of times it can discover unknowns like things or markets or segments you wouldn’t have thought to go after, which is where it really comes into play in advertising, like where you can set a budget, and you can develop the creative, but when you let the AI actually run, and potentially discover new markets, it may actually find buying signals that you didn’t know existed.

Paul Roetzer: So, you know, I think a lot about this with our conference. Like that’s probably the most scientific approach I’ve ever taken to anything in marketing is how to sell conference tickets, because we’re thinking about all of these different things that could be signals that someone would likely purchase a ticket. And so there are definitely things that come across or I would see people buy, it’s like, why did they bought, like, I would’ve never guessed that they would have been a buyer of a conference ticket.

Paul Roetzer: And so as marketers, I think one of the first things you always wanna do is accept like we kinda suck at strategy. Like it’s really hard. I’ve been doing this for 22 years, and I’m the first to admit like, there’s things I think are home runs like, no doubt, this is gonna work, and then it falls flat and you’re like, what was that all about? And then there’s things where I’ll try something, and be like, oh, it might work, and all of a sudden, like that was the thing. And so with AI, you have that additional layer of machine intelligence helping us, you know, humans try and solve what is very complex. The more data you have, the harder it gets to figure out the right strategies.

Tessa Burg: Yeah. I love that ’cause it also validates the point you said earlier, like just get started, and try, and be open to those learnings, ’cause that ultimately is what’s going to drive the most success. I feel like a lot of people kind of get caught up in that overthinking, and if you don’t start —

Paul Roetzer: Or thinking you have to have all the answers. Like it’s not what marketing is, it’s trial and error, and it’s more so now than ever.

Tessa Burg: So let’s move into the next stage of the sales funnel. This is demo, engagements, how can AI help marketers create content, create sales material that really resonates, and moves those prospects through the funnel?

Paul Roetzer: So one thing that, you know, kind of nearly came to mind when I saw this one as conversational AI, like that’s a term you’ll hear quite a bit, especially in the enterprise world, a lot of the big companies, specifically in retail, and e-commerce that are making major investments in AI, are putting it into conversational AI. And by that, what I largely mean is not a chat bot, that, you know, you go to a site, and you just assume like, oh, okay, this chat bot’s stupid, like I’m asking questions, it’s not gonna have the answer, I’m gonna say, let me talk to a human, these are conversational agents that actually understand questions. It’s kind of like when you go to Google, and you search for something, and it’s like a broken sentence or part of a question, but Google, somehow actually understands the intent of your question and delivers a result to you, that’s like, yeah, that’s what I needed.

Paul Roetzer: Think of that applied to conversational agents on websites where I can, it knows the pages I’ve looked at, it can look at the CRM record, and know if I’ve been there before or taken previous actions if I’m a customer or not a customer, and it can actually adapt its conversation with me based on all that information. And then as I’m asking questions, it’s learning from that.

Paul Roetzer: And so not only does the machine learn, but us marketers and salespeople can learn from those interactions to better develop content and communications based on what are people actually asking for. Because, you know, if you think about it, what do we know about someone that comes to our site? It’s maybe with the keywords they searched for, maybe the referral place they came from. But if we don’t have a record of them previously, we don’t know a heck of a lot about them.

Paul Roetzer: So if you develop a knowledge assistant where people can ask questions or like tell you what they’re looking for, think of all the information you can start learning about that visitor who may still be anonymous at that point, but you start to get a rich data set. Now, again, if you have a site that’s getting 100,000 visitors a month, you may have a thousand engagements, you know, conversations, no human is gonna read through a thousand conversations, and like extract meaning, but AI can read through almost an infinite amount of conversations, and then use language processing to extract what are they actually asking for, you know, we take the whole dataset, what are the commonalities here, like what are the things that we can apply to create something of value?

Paul Roetzer: So if 75 people asked relatively the same question last month, well there’s hey, demand team, can you create a download for this? ‘Cause this is a question we’re getting daily now. And now the conversational agent, instead of, you know, not having an answer can say, “Yeah, here, download this, “this answers your question.” So that’s where you start to inform, you know, the demand generation side with the conversational side, and you know, the qualification.

Tessa Burg: Do you have an example of another tool that you think is doing that really well today?

Paul Roetzer: Frase.io, it’s F-R-A-S-E.io is one we’ve used, that’s a knowledge assistant, so that’s what it does, the beauty of that one, one, it’s affordable. Two, it learns from your existing site. There’s no training needed on day one. So all you do is you plug it in, and it actually then consumes everything on your site, as it’s knowledge-based to start. Now, you can monitor the responses it gives, and you can tell it, you know, no that wasn’t a good do this next time, but it continually learns from that human training and input, as well, so that’s a really cool one, Drift is kind of a leader in the conversational AI space when you’re looking for the more robust solution, LivePerson would be another like enterprise-grade conversational AI solution, but Drift is the one I’m most familiar with, and they’re making major investments in AI.

Tessa Burg: That is awesome. I’m taking notes so that we can go with our clients, and look at some of these tools, as well. So our last phase of the funnel is closing. Can you share some examples of how AI can help shorten the sales cycle?

Paul Roetzer: So again, a lot of companies are making investments in this space and that is language, understanding, and generation applied to sales emails. So think of it as when you’re going and writing an email that you almost have, like many of us use Grammarly or a similar tool that assesses our writing as we’re writing, and in some cases can even score the quality of your writing, think about that kind of technology applied to your sales emails, where it’s actually recommending ways to improve the email as you’re writing it, potentially even scoring the email, or in some cases writing it for you.

Paul Roetzer: So if you think about how Google smart compose finishes your sentences for you, that’s made possible by AI. That same technology is starting to be integrated into sales technology, where they’ll actually start finishing your sentences or writing paragraphs for you in real time, but that can also then be applied to developing templates. So one is the communications, and like the cadence, and the actual language you use, two is being able to monitor opportunities to flag for you when actions should be taken.

Paul Roetzer: So, you know, if you think about you’re just going in everyday and checking your pipeline, and setting your reminders to follow up in three days, or whatever that is, you may have an AI, for example, that learns that Jane opens her emails on Saturdays at 9:00 PM, like it’s the most common time Jane opens emails. So rather than sending your email Friday at 9:00 AM, it might tell you, hold on, don’t send this email, she’s opened her last three emails on Saturdays, actually. So let’s schedule this for a Saturday. Something a human is never going to do at scale. You’re not going to drill at an individual level, and personalize those communications like that, that’s where AI in the next three to five years will be omnipresent. It’ll just be integrated into all salespeople’s workflows, and kind of think of it as an assistant there for you to optimize your opportunities for conversion.

Tessa Burg: I love that example, and I feel like this conversation has really had it all, you know, we have successfully, or really you, I didn’t do anything, but made AI more accessible. I think a lot of listeners have realized AI is already in their life, and so it really is about, you know, using the resources to make it more accessible, and easier to understand in your organization so that you can start to realize some of the value of prediction, of learning, upscale that really only AI can do, and that just isn’t possible through people power. Paul, if people wanna get ahold of you and reach you, where can they find you?

Paul Roetzer: I’m pretty good on LinkedIn, I mean, Twitter, I’m pretty active, it’s just at Paul Roetzer. but LinkedIn for sure, same, just search my name on LinkedIn, the AI will find me for you, but I’m pretty good on responding to messages, and connection requests there, and then my email is just [email protected].

Tessa Burg: Awesome, well thank you much —

Paul Roetzer: I’m not really active on TikTok.

Tessa Burg: No? You’re not active on TikTok? Ah it’s surprising.

Paul Roetzer: Conversation for another time, but …

Tessa Burg: All right, I am definitely not a fan of TikTok’s AI, because I am very easily sucked in, and it has me pegged. I really love dancing videos. Well, thank you so much for being our guest. If you wanna hear more episodes from Tenlo Radio, and Leader Generation, visit tenlo.com, and click on podcast, we’ll have this episode, as well as the list of tools that Paul has shared, and a link to marketingaiinstitute.com, and the spelling of his last name, which doesn’t sound like Roetzer, will all be there. So check it out, and we will see you again in a couple of weeks, we have recently changed the show to releasing every two weeks. So you will be hearing from myself or Cheryl Boehm, who is our head of content and copywriting. Thanks.

Paul Roetzer: Thank you.