Challenges Marketers are Facing Implementing Generative AI

Mike Boyle: Paul, could you share any challenges that you maybe have come across with other B2B marketers, things that they’re facing while implementing generative AI into their day-to-day content prep?

Paul Slack: I touched on one of them a little bit earlier when we kind of kicked off the podcast. But I think one of the biggest challenges is just the overwhelm that marketers, and I think anybody that’s trying to leverage AI, is feeling right now from some of the things we talked about with how many newsletters are out there, how many tools are showing up on the scene every day, and then how full our plates already are. And so just trying to get our heads around it has been a big challenge, and I think I feel it because I talk to folks every day, either on social media or in meetings, and I just sense that this is really kind of an overwhelming thing. Another challenge that I’m seeing a lot of is, and, you touched on it a little bit, Mike, with the prompting and prompt engineering, and happy to dig on that if you want a little bit on how to get some good output. But the challenge is they’re not telling AI to do the right things and so they’re not getting the right information back. And one thing I try to communicate to folks is you need to think of AI like a raw material, like wood or a Lego. And so I think people approach AI, it knows everything, but it needs very clear directions. And so you can’t just go to a log and say, be a table, right? It’s not going to do that, right? And you can’t just go to Legos and just expect it to create something amazing. We have to have instructions and we have to have a plan. And so one of the things that I communicate to folks is to overcome these types of challenges. Realize that AI is really one piece of a three-step process that’s needed to get the most out of it. One, you have to have the right tool. So go pick the right tool for your business. In our case, that’s a suite of tools, but, start with one as we talked about earlier. But you have to have the right tool. So the tool that’s going to get you closest to the type of work that you’re trying to do, that’s step one. But step two is you have to have people that are trained on how to get the value out of the tool. So that’s step two. And then step three is you have to have methods and prompts and best practices on what it is that you’re trying to get accomplished. And so if I take an inexperienced person, remember, think of AI like a tool, a chainsaw as well. And I just say, go cut down some trees, they’re going to have a hard time. But if I train them on how to use a chainsaw and then I walk them through the steps that are, hey, when you’re trying to cut down a tree, these are the steps you need to take to get that tree down. so that you don’t have any challenges when you’re doing that, that’s when you’re going to have success. And so where people are struggling is they’re jumping into the deep end. they’re asking it to do broad things and then they’re getting output that they’re like, well, this isn’t helpful. This is a waste of time. And then, they become more frustrated.

Mike Boyle: In my world, in Salesforce, there’s great tools for AI. They [Salesforce] have just a whole plethora of great GPT tools now. And one area that’s most important to businesses always is customer support and engagement…Talk to me about how AI has improved customer support and engagement for B2B marketers.

Paul Slack: You know, Mike, I think that you’re touching on something that we’re just at the very beginning of when it comes to AI, and when we think about 2024, I feel like from my seat this year was very much about content and leveraging AI for marketing things. But in 2024, I do think you’re going to really see customer support get some pretty cool legs when it comes to AI. And the reason why I say that is, yes, we can absolutely use generative AI to help, provide, build chatbots, and get some immediate feedback and answers to questions way faster than trying to Google something or go into a knowledge base or whatever. It’s also very good at translating languages. And so if you have content that’s in English and you need it to be in other languages, or vice versa, to help mitigate the calls that you’re getting and customer support, my goodness. ChatGPT and AI is very good at those kinds of things. In 2024, though, with some announcements that just happened this month, with ChatGPT announcing GPTs, which are basically creating your own version of ChatGPT, where you train it on things that are just relevant to you, that’s going to have a huge impact on customer support, moving forward. So imagine, Mike literally having everything that you’ve ever written or every video that’s ever been created or anything about your product or service that’s ever been published in one place, and then you build a GPT, A, mini chat GPT that knows all of those things. And so now think about your employees, think about customer support, think about chats, interactive chats on your website. It can all touch on that knowledge base and not give what’s known as hallucinations in the world of AI, but actually give very relevant feedback, either to your customer service team or directly to customers as they interact with you online or whatever the case may be. So, big impacts are coming for sure.

Mike Boyle: You’ve touched a little bit, Paul, on, content generation. I’m curious about how AI can help marketers with content generation. Could you expound just a little bit more for me on how generative AI will assist B2B marketers in their content generation?

Paul Slack: There is so much there, it’s incredible at just ideation. So, let’s start with what you mentioned earlier, Mike, about asking the right questions. And so, there’s two pieces to really training AI. It’s one, understanding, prompt engineering. And Mike, you and I talked before the show. We’ll drop some links to resources in the show notes, to some great tools that we’ve made available. One of them is a prompt engineering guide that teaches what’s called a race framework. For the sake of time, I won’t walk you through that, but it’ll teach you how to talk to AI the right way to get good outputs. The other thing that you need are seasoning documents. So, seasoning documents just simply are your persona, your voice and tone, some key things about your business that you can have documented and shared. This kind of goes back to that three-legged stool where we have to have the right tool, we have to have people trained on the tool, then we need to have the right processes. Well, seasoning documents are part of the process, right? It’s something that you can use different people in your organization will be able to use to quickly train up AI to get the outputs that you’re looking for. Now, once you have those things in place, then you can go and do some really cool stuff like data-driven ideation. And so one of the things that we like to do, Mike, is take the content on our website, take our key phrase list, of all the words that we want to be found for. take our persona and all the things that they care about, mash those things together inside of AI, and then ask it to tell us where are the gaps. Because one of our goals as marketers is we want full topical coverage on the key phrases that we want to be found for. That helps with SEO. It also helps us communicate to our customers. And so these are things that Mike would almost be impossible to do with a human. But to put all that information in AI and then go to it and say, hey, based on the information that you have, where are the gaps? Like what topic areas are we not talking enough about or not talking about at all? And it will tell you that. And so, ideation is always the beginning of good content marketing. And so the other thing we can do is repurpose content. We do that a ton. For example, we do a monthly event, it’s called the Demand Gen Jam session, where we’ll teach on a marketing topic or have a guest that’s an expert on a marketing topic. Mike, we take those transcripts, which are thousands of words long, because they’re about an hour-long discussion. And then we can upload that information into an AI tool and then have it again. Ideate, tell us, hey, based on this content that you’re based on this transcript, what are some good topics that we should be writing about? It’ll tell you that, but then you can also say, great, you gave me five. I like one and three. Those are really good. Using this transcript, write an outline for a blog for topic one and topic three, and it’ll create an outline, and then you can actually begin to use it as an AI assistant, to help you flush out and really create some remarkable content. Some of the tools are really good for research. Now, ChatGBT out of the box doesn’t connect to the web, although you can add plugins and accomplish that. We love being able to go and do research. It’s like Google on steroids, Mike. So we can go in and once we have an AI talking to the Internet, we can do some pretty incredible research. Like imagine writing a blog using a transcript, but then saying, you know, can you find some stats about this point? And then it’ll go out and bring back stats to you and you can incorporate those easy into your content marketing. So the reality to all of this is it’s just making us better marketers and copywriters, and it’s making us better and cutting time at the same time. It also allows, from my perspective, the ability to upskill my team. And so I can have people that aren’t necessarily great copywriters, do a lot of the content creation work, still, needing to go through an editorial review and things like that, where I wouldn’t necessarily, before chat, GPT or AI, given them tasks like that, because maybe it was a little outside of their skill set, but now AI has been able to allow me to upskill them and get them to do work that normally they wouldn’t be able to do. And the last thing is just analyzing data, right? So it’s kind of one of the first points of doing data-driven ideation. But I want to be able to go in and analyze, what content is working for me, what content is not working for me. And AI can do that again, much better than a human could just because of its ability to consume massive amounts of information.

Mike Boyle: I want to, very briefly, lead gen and conversion rates. But before we get there, there was one question that I wanted to ask you about, and I know you do this as the owner of your business, and that’s hiring people, new hiring. I’m beginning to read things about new hire training and how generative AI can help in that area. What have you experienced, if anything at all?

Paul Slack: One of the things that we are always trying to stay on top of the leading edge when it comes to some of these things. But in terms of just new hire training, some of the things that we’re looking at, and I know some of the things that other companies are starting to explore is like taking their training manuals and their onboarding documents and converting them either into a knowledge base that can be, available to them in an AI experience, a generative AI experience. One of the cool things that I’ve started to see people do is take their text-driven, training materials and converting those into AI videos. And we haven’t talked about this too much yet, Mike, but you can create video avatars that look like human even. Mike, we could train an AI to look just like me or look just like you, with enough, uploading of images and things like that, and then turn in, create a virtual trainer that looks like me, looks like you, or could just be more of a generic, avatar type person, and then create video content versus text-based content. And one thing that I’ve always heard in training is that when you can see it, your retention is a little bit better. And also, people prefer to learn in segments of information. And so typically training manuals, we all know this, they can be volumes long and it’s hard to absorb all that. But if you can take that content and break it into five-minute little video vignettes and have a video avatar produce it, what it’ll look like to the new employee is just a trainer that’s giving them information in a video format versus them having to read the content. I feel like there’s going to be a lot of that happening in the near future. And then, going back to these GPTs, this announcement that chat GPT made earlier, this, know, taking a GPT and training it specifically on your sops or specifically on, the things that you do that make you unique and then making that available to new hires as they’re trying to learn and get up to speed, I think is going to really shorten that learning curve when people are trying to get started in their jobs. And one thing I’ve already touched on is just that upskilling of workers. I just feel like one great thing that AI is going to be able to do in the new hire process is you’ll be able to take an intern that you just brought into, an entry-level position and get them doing some pretty high-level work, especially in my world, which is the marketing world, quickly. So those are some areas that I really think AI is going to help with training new folks.

How AI Enhances Lead Generation and Conversion Rates

Mike Boyle: Let’s get into just briefly about lead gen and conversion rates. I know these are two topics that are big in my world for our customers at Ad Victoriam Solutions and I know for your customers at Vende Digital. So talk to me about how leveraging AI can enhance those two things, lead gen and conversion rates.

Paul Slack: Yeah, I’ll tell you, there’s some amazing opportunities now. And in my mind, again, from a marketer’s point of view, what this is allowing us to do is tap into the kind of work we would love to do but we’ve never had time to do. And so when I think about it, let’s just take lead generation first. So from a lead generation perspective, what we can do with AI is really hyper-personalize our information and that’s going to be key. I’m going to try to paint a picture for you here, but let’s start with you’ve got really good content on your website. It speaks about your services and the value proposition. Imagine being able to go in and repurpose all of that content to speak to all of your key buyer personas across all the different industries that you’re trying to serve. And then you create that content and you put it on your website as in a content hub. And so now you have your base content on your website that you’ve always had, but now you have a specific content hub that’s going to meet the needs of a particular industry persona or a buyer persona. And you use your marketing to drive people to these very personalized, hyper-personalized experiences. And why that’s going to help in lead gen is when you’re using a tool like Salesforce or any of the big tools that are out there, you’re going to be able to monitor consumption, ah, of that information. And so in today’s world, I’m going to really kind of oversimplify this. We’ve got an offer, we’re going to run some campaigns to generate leads to that offer, and then we’re going to capture some leads as best we can. And there’s nothing wrong with that. I call that kind of a bottom-of-the-funnel strategy. But what we’re going to be able to do because of this hyper-personalization and creating content hubs is we’re going to be able to run some campaigns that are going to drive interest and intent, type data to these content hubs, and then use that engagement data as a way of modifying lead score and kicking off very hyper-personalized campaigns for those individuals that we just wouldn’t have the ability to do in the past. And so where we’re going to go with improving lead generation, and you can already do this today, is really driving personalized experiences for your buyer across their, you know, statistically, when we’re touching them and serving them at every stage of their buyer journey. Mike, our conversion rates increase according to LinkedIn sixfold. And so when we’re focusing on just the bottom of the funnel, hey, here’s an ebook, here’s a form, would you please fill it out? That’s bottom of the funnel. But when we’re doing full customer journey marketing and giving them personalized experiences, we get six times the number of leads through that process. So that’s one way that we can improve lead generation, is just through hyper-personalized content and leveraging it the right way. You can’t just personalize the content, but you’ve got to use it to create personalized experiences.

AI Helps with Ad Optimization and Analysis

Paul Slack: Another area is in ad optimization. And I know in my industry, people freaked out over this at first, but I’m a big believer in adopting new technologies and figuring out how to leverage them, and passing the value on to our clients. And so when AI came out, I immediately wanted to see, how can I use this to benefit our ad optimization. And so there’s tools out there, AI tools that you can buy that will help you optimize your ads. We’re not using those at Vende. What we are doing, though, Mike, is we’re dumping all of our campaign data into AI and having it analyze for us what ads, what creative, what content, what audiences are performing the best. And we’re learning things way faster and getting insights we could never see on our own by having AI help us with that. And so the output of that or the benefit is we’re now putting what we call recipes, Mike, right audience, right message, right creative, in market faster, that are converting better because of using AI to help us with ad optimization and analysis. The last one I’ll touch on is really in just auditing your CRO experience. And so I touched on this a little bit earlier. You can take your landing page, you can season AI with your personas and your information, and then ask it to tell you what’s missing on this landing page, and it will tell you. And then you can go and update and optimize your content based on what your persona or the AI is telling you is missing. You can also do really good competitive analysis the same way. One thing I’m really enjoying, Mike, is, you know, with ChatGPT Plus, which is the paid version, you can screenshot now your entire landing page, and you can just throw the screenshot in and then season it with your persona. And, you know, one of our personas is Samantha. You’re Samantha. What’s missing from this page? And I don’t have to type anything else. I can just literally screenshot a picture of a landing page or one of our client’s landing pages and it’ll start telling me what’s missing.

Paul Slack: Now, getting back to good prompts, you have to tell it what you. One of the first things you do in prompt engineering is you tell it the role. And so, like, if I were going to be improving my CRO on a landing page, I would say you are a CRO expert in this industry. To start, that’s kind of helping it understand the context. And then you could put in a landing page screenshot or something like that, and get some really interesting information back to help you tune your landing pages. But those are some big ones.

Mike Boyle: We started this conversation off, actually, you brought it up. You said, Mike, can you imagine what we were doing this time last year? We didn’t even know about ChatGPT. We didn’t know how it was going to explode. Well, here we are, almost twelve months later. It’s been one heck of a 2023 in terms of AI and GPT and everything involved. Okay, we’re here, we’re getting knee-deep into it. What’s 2024 going to bring us in this world, Paul?

Paul Slack: Mike, I’m going to make a prediction right now, so hopefully you’ll invite me back this time next year. I’d love to.

Mike Boyle: Count on it… Count on it.

Paul Slack: My guess is, we’ll be having this conversation this time next year. And the things we’re talking about today, and even the things we’re about to talk about are going to feel really antiquated, you know what I mean?

Mike Boyle: We’re going to be yawning.

Paul Slack: Oh my gosh. So take all that with a grain of salt because this stuff is changing so hyper-fast, it’s unbelievable. But when I think about where we are today and where I think we’ll be this time next year is, more efficiencies gained, like when you start working with AI. Yes, it’s a big-time saver. but there’s a lot of things that you’re having to do to season AI or to put certain prompts in and somewhat feel laborious like in order to convert a transcript into blog content, I’ve got to go to one tool and upload a video and have the AI of that tool turn that into a transcript. And then I’ve got to download that, I’ve got to take it over to another tool, I’ve got to upload it into another tool and start interacting with it to get ideation or outlines for blogs or whatever. And so while all that’s incredible, didn’t have it a year ago, there’s still a lot of steps involved with that process. And so I really do believe a lot of efficiencies are going to occur next year and some of that stuff’s just going to get super automated. Imagine like right now we’re doing this podcast episode. I can imagine a year from now, as soon as you hit stop that everything’s just going to happen for you. You’re going to have the transcript, it’s going to give you the show notes, it’s going to pull out some of the key, snippets, and highlights, from the session. It’s going to do so much more just based on some of those inefficiencies that I just talked about. So that’s one thing. Another one is we talked about personalization, but I really do think that we’re going to see that at scale. So today it’s still very laborious. It’s awesome. We can take our legacy content that’s performed well, we can hyper-personalize it, we can create content hubs around that. But imagine in a year from now, or less than a year from now, Mike, uploading a blog, it already knows all the personas that matter to you, and immediately it’s spitting out and creating those content hubs without hardly any oversight from a human. I do believe that’s going to happen a lot. So today it’s amazing, but it’s still laborious tomorrow, literally, I think you push publish on a blog and immediately you’re going to get ten other variations of that blog that can be used for hyper-personalization and customer segmentation. So that’s going to happen for sure. And then predictive analytics. So today there is predictive analytics. most of that’s baked inside of tools like Salesforce, for example. but I think we’re going to be as marketers and sellers, and operators of businesses. We’re going to be able to leverage predictive analytics on our own, which all predictive analytics means is looking at your data and then helping you make intelligent decisions. About the future. And so today as businesses, we’re looking at data and we’re having to figure that out on our own. I think we’re going to see way more automated predictive analytics that’s going to help us make better decisions in the future.

AI Will Automate Many Jobs

Paul Slack: The last one I’ll touch on, Mike, are these GPTs? I’ve mentioned GPTs like three times I think in today’s podcast – they’re brand new, just off the presses. And Mike, I can send you a link, we can drop in the show notes about Open AI’s announcement around that, but these now, imagine having your own chat GPT that is trained to do a specific function in your business or provide specific information to help your team or your customers. That’s going to explode and then what’s going to come from that really are AI agents. And so by this time next year, Mike, I predict that not only are we going to get outputs like we’re getting today, but those outputs will also be instruction sets that then AI will go and do for us. So imagine taking Zapier, which is a very popular automation tool that connects different systems together, but merging that kind of capability, with AI, so that when you go and ideate something, it’ll just go and create multiple steps for you so that you’re not having to do the human steps that are involved today. So these AI agents, you’ll hear a lot more about that down the road. We’re not quite there yet, but that’s going to be the future. When the robot can actually go, it’ll tell you what to do, but then it can actually go do it for you. So that’s going to be pretty exciting.

Mike Boyle: Yeah, exciting is the word. And as Paul has mentioned, and I’ve mentioned a couple of times throughout the podcast, there are going to be a boatload of links that we’re going to put in this show’s notes. Paul has been gracious enough to put together some really great AI tools and tips and I’m going to have links to Paul’s things, in the show’s notes, in addition to an ebook that AdVic® just recently put out outlining all the benefits of Salesforce’s new GPT tools. So make sure you look for all of those, great, links with tons of information, in this show’s notes. Paul, just a big old thank you for talking to us today about AI and marketing. I hope you never get tired of me calling on you, Mike.

Paul Slack: I love doing this show every year and look forward to connecting with everyone next year. And definitely, please take advantage of those resources. My team’s done a really good job of putting together some toolkits that I think will help you guys if you’re really interested in learning more about AI. And they’re great, little resources and, hopefully, you get some value out of them.

Mike Boyle: One quick thing that I wanted to mention to our audience because they would really be interested in this if they don’t already know about it. But you also head, on LinkedIn, a group called, Demand Gen Jammers. How can people get involved in that and participate in some of the live events that you do?

Paul Slack: Yeah, absolutely. So we do have a LinkedIn group. It’s called Demand Gen Jammers. And, I’ll make sure that we have links to that, too, so you guys can join it on LinkedIn. It’s just a place where we can help each other, stay on top of everything that’s going on in the world of B2B. Very heavy emphasis on B2B marketing. A lot of stuff on AI, as you might imagine, because that’s impacting our jobs, quite a bit. And so it’s a community for sure. And then once a month we have these one-hour sessions. They’re at [12:00] Central time, about the middle of the month, on Wednesdays, each month. And we have anywhere from 100 or so people that show up, and they’re pretty informal. Mike, I know you’ve been to a bunch of them, but we just wrap on a topic of the day, like, we’re about to have one tomorrow. It’s going to be on AI use cases for B2B marketers. And, next month’s, probably about the time they’re hearing this podcast, the one coming up will be, all about developing a 2024 strategy… and what are some of the big trends that we’re seeing in the world of just B2B in general? And so they’re topics like that. And sometimes I’ll invite really smart people to come and share their perspective. And really, the idea is just to have a conversation. And so would love to have everybody that’s listening in, that’s interested in staying up to date on B two B. come join us. We’d love to have you.

Mike Boyle: And, we love having you, Paul, thanks again, for being with us. If you want to learn a little bit more quickly right now about Paul and what his team does at Vende, digital, just visit Vende Digital. That’s V-e-n-d-eDigital.com. Talk to you again soon, Paul, thanks. Appreciate it.

Paul Slack: Thanks, Mike. Enjoyed it.

Mike Boyle: And if everyone listening has enjoyed the podcast, please let your friends know we’re available on any podcast channel, including now on YouTube music. And make sure you download either that Android or iOS app. And, you’ll be able to handily find this show there as well… I’m Mike Boyle from Ad Victoriam Solutions. As always, thank you for joining us for the “Salesforce Simplified” podcast. Our next episode is just around the corner.

Speaker A: You’ve been listening to “Salesforce Simplified,” the podcast from Ad Victoriam Solutions.