How to Leverage AI in Marketing: What We’ve Learned | Part 1

AI is rapidly transforming marketing by automating tasks, enhancing creativity, and optimizing workflows. It’s no longer just for tech companies—as B2B marketers, we all can now easily integrate AI into our strategies to drive efficiency and innovation. Many marketers are still in the experimental phase, so in this episode, we discuss what we’ve been experimenting with when it comes to AI here at GrowthMode Marketing, what’s working for us, and what’s not.

(00:00:20) The potential of AI to transform marketing strategies
(00:02:21) Experimenting with ChatGPT 
(00:03:16) Brainstorming ideas with AI
(00:05:38) Providing detailed information to AI for better output
(00:07:42) Creating content outlines with AI
(00:10:00) Using transcripts to create social media posts
(00:14:41) Cleaning up transcripts
(00:15:59) Identifying areas in video podcasts that require editing
(00:16:06) Proofreading and receiving detailed recommendations for writing improvements
(00:17:39) Summarizing meeting notes and extracting insights
(00:19:03) Identifying key points of reports and articles
(00:19:22) Formatting data into tables, lists and notes
(00:23:12) Identifying inefficient tasks that AI can help simplify 

The Demand Gen Fix is hosted by Deanna Shimota, CEO of GrowthMode Marketing. Listen to our team of marketing experts and featured guests drop knowledge on how HR tech companies can maximize the success of their marketing for today’s B2B prospects. Learn more at www.growthmodemarketing.com.

Transcript

Jenni Geiser 00:00:01  Hey, everybody, it's Jenni from GrowthMode Marketing. You're listening to The Demand Gen Fix, the podcast where our team of GrowthModers and our guests discuss the ins and outs of demand generation, and why we believe it's the key to long-term sustainable growth, especially in the HR tech industry.

Deanna Shimota 00:00:20  Hello. Hello, Deanna and Greg back for another episode of the The Demand Gen Fix podcast. And today we are talking about how to leverage AI in marketing, and more specifically, what we've learned here at GrowthMode Marketing as we've experimented with it. This is a multi-part podcast series. We had one episode already. After this, we'll have 1 or 2 more episodes, depending on how much content we get through today, so keep tuning in to learn more about AI and marketing. AI is rapidly transforming marketing by automating tasks, enhancing creativity, and optimizing workflows. It's no longer just for tech companies, and as B2B marketers, we can all now easily integrate AI into our strategies to drive efficiency and innovation. I know that this was scary for some marketers out there.

Deanna Shimota 00:01:18  I've heard people say I don't want it to replace my job. This is too much. It's hard to figure out. We get it. We've been there. In the last episode, which was episode 69 when we kicked off the series on talking about AI. We shared tips on how to find AI resources and start experimenting with it so you can jump in and start to learn.

Greg Padley 00:01:45  So before we dive into what's working for us, it's worth noting that AI adoption is still evolving. As Deanna said, many marketers are afraid. Many marketers are still experimental phase of things. So in this episode, we'll talk about what we've been playing around with and experimenting with here at GrowthMode Marketing and what's working for us. It's funny, just the other day I saw a report that came out from Grammarly and it said 52% of knowledge workers didn't know anything, didn't know anything about AI, and didn't know where to start. If you're feeling like you don't know where to start. You're not alone by any stretch.

Deanna Shimota 00:02:21  Nope.

Deanna Shimota 00:02:21  Not at all. So at GrowthMode Marketing, we've been experimenting mainly with the paid version of ChatGPT, which is 4.0. I think they now have a beta version of Another Step Forward. I've started playing with that. Great. You probably have two. So everything we're going to talk about is around our experience, mostly with ChatGPT. I've played around with Claude. I know you've played around with some different tools as well, Greg, but we're going to share with you some of the things we've done that we've found ways to work it into our marketing Efforts and work. This is not something that we have started to use with clients outright, yet we're doing a lot of experimenting internally to figure out how to make it work and how to make us more efficient without it having an impact on our clients. So with that, let's jump in.

Greg Padley 00:03:16  Yeah, I find one of the things that I use it for a lot is brainstorming. Just coming up with ideas just on the fly. You could come up with different strategies, different campaign ideas, different content themes, like we used to think of product names and titles of articles.

Greg Padley 00:03:33  So it just helps to. It's like having another person to a sounding board, just like brainstorming is right, except you don't actually have another person there using the AI, right?

Deanna Shimota 00:03:43  I have used it for brainstorming as well, and I will say it's not like you can just go into something and say, hey, give me ideas for an email campaign that we can run here at GrowthMode Marketing. When you do that, you get garbage. And I gotta say, honestly, when we first started experimenting with, I and other people were telling me like, oh my gosh, it's saving us so much time with some of our marketing processes. I would go in and play with ChatGPT, and I felt like what I was getting back was not usable. Part of it was I was like, hey, write me a really good email campaign to do outreach to prospects. But that's very vague feedback, quite frankly, on my part. I've learned a lot since playing with GPT and figuring out how I can use it and what it does well and what it doesn't do well, but it didn't necessarily work real well for me in the beginning, no matter what I tried to do with it.

Deanna Shimota 00:04:44  and so I was like, this is garbage. I don't know why other people are raving about it. I was trying to use it to help me brainstorm topic ideas for social media, for example, and get help with actually writing the social media post, because I was doing a lot of posts on LinkedIn for visibility, for growth mode, and I wasn't having a lot of success with it, I would get it, not like it, and end up go and write from scratch anyway, right? So I think with the brainstorming, one of the things that works well with it is really like knowing exactly what your brainstorming for and feeding it information first. And I don't know what your experience is on this rig, but I've found like just going in and asking it a question you're not going to get.

Greg Padley 00:05:38  Yeah, you're absolutely right. When I started talking, I just jumped into talking about what I was using it for without saying what you just said, which is absolutely right. I think we said it in the last podcast.

Greg Padley 00:05:47  It's like your AI is your intern, and if you don't give the intern good information to start with in good directions. You're not going to get anything good out of your intern. And it's the same thing with the AI. Like, you really need to give it some details and some information and some criteria. Who's your target audience and all that kind of things that you would pass on to even a creative person, right? Like they would need to know that to create something. So it's similar to that. You just have to have a good set of information to start out right.

Deanna Shimota 00:06:17  It's almost like you still need that creative brief. To your point, if you're bringing in another person on the project, you would have to fill them in on the details. You have to fill ChatGPT in on the details to write if you're going to get good stuff back. And from a brainstorming standpoint, I don't feel like I've ever gotten anything that came back that I'm like, perfect. I don't need to do any more work here.

Deanna Shimota 00:06:40  I've got my list of ideas in the winners right there. It doesn't work that way. I wish it did at times because it would be a huge time saver, right? But it has some good ideas, right? And I have some ideas where you're were like, that's not going to get used, but you need that human oversight to edit it and continue to ask questions and maybe go back in and edit the prompts that you did to have it, redo it until you get it to a point where I would say at times it saves you maybe 50% of the work of doing it from scratch. It's never 100%.

Greg Padley 00:07:16  Right. And I think that's brainstorming is one of those things that it's like a human thing, right? It's like creative kind of process. It's people getting together and spinning things around and coming up with an idea. I don't know that you can totally give that over to a machine to do that for you.

Deanna Shimota 00:07:31  It's more like it tosses things at you to help you keep evolving your thinking, and pull little threads in your mind to come up with the end ideas, right? Yeah.

Deanna Shimota 00:07:42  Another way that we have used it is to prepare an outline, like a content outline, and to do this. What we've had to do is take like all of growth modes, content, for example, like our marketing content. And we actually built out a GPT. So a custom GPT where we're feeding the information in this is all content we've previously written. So it's like creating a database of knowledge around our content and topics we typically talk about and how we say things, and then going to it and asking it. Create an outline for an article on XYZ topic and it will come back with an outline recommendation. Now again, I find first time out I'm not usually perfect. Let's run with it. You have to keep feeding it some information and getting it to refine it to a point where again, not 100%, but you may have something that's 50 to 70% where you need it to be, that you can then take and start to massage it to create a really good outline for the content.

Greg Padley 00:08:58  Yeah, I agree, and sometimes it's not that it's necessarily giving you the wrong answer or I don't really want to talk about that, but it doesn't know that you don't want to talk about that, but it's coming back based on all the information you've given it.

Greg Padley 00:09:10  And I don't want to say best guess, but. Right. It's like putting the pieces of data together to come back with the answer. and it could be the answer for somebody else. It could not be the answer for you because you don't want to write about that time.

Deanna Shimota 00:09:23  Right? And I would say a limitation with building your ChatGPT or your GPT, and then coming up with content ideas out of it, it's not going to come up with a new and novel content ideas because it's basing it off of your historical content. Just I'll have to keep that in mind as you're building it. You want to create content, for example, social media posts based on all of the work you've done in the past, are you trying to come up with a new and novel idea? And if you're trying to come up with a new and novel idea, a database of all your past content probably isn't going to produce that.

Greg Padley 00:10:00  And with the social media posts, some of them, if it's a post about a podcast that we've done, you can take the transcript and start to create a post out of that again, comes back and it's close, but it's not quite there sometimes.

Greg Padley 00:10:13  And you do have to give it some direction on what kind of tone and voice you want it to have, that it's similar to what all the data that you've put into all the content you already put in there.

Deanna Shimota 00:10:24  Yeah. So I have in the past because I've posted a lot of things on LinkedIn, for example, that were thoughts that came out of my head that I wanted to share with my audience, but it's a lot of work to do that on the regular. And I was like, gosh, it would be wonderful if ChatGPT could just look at my LinkedIn profile and all of my LinkedIn posts and come up with some new content. For me, that's never worked for me. One note though ChatGPT doesn't really want to read the website and pull that information, so there's a little bit of a limitation or maybe user lack of experience there to be able to do that. And I agree it can get you the framework or like partially there for writing a post, but I never get post back where I'm like, sweet, I'm going to use this as is because I'm like, okay, I told it to write in the tone and style of Deanna Shimota at GrowthMode Marketing.

Deanna Shimota 00:11:18  I gave it a link to my LinkedIn. Like I tried to feed all those things to get it there, and it doesn't quite get there, but it can certainly help you get closer to there. And so if you're somebody who it's easier to edit something than it is to start from scratch, which I know for a lot of people, reacting and massaging something can be easier than putting the initial words on the paper. Then it's a great tool to use ChatGPT for that. And I don't know, Greg, I know you've done a lot of social posts and leveraged ChatGPT. Maybe you've had a different experience than me and trying to use that.

Greg Padley 00:11:56  No, I would say it's similar, but what I've done with that is I keep going back to the same chat. If I'm posting some of our growth mode stuff, I just keep going back to that same chat and then I'll say, before you do anything, go back and read this chat that we've been having, and then I'll give it the next, and then read this piece and write a social post about it.

Greg Padley 00:12:22  So I try to remind it, go back and read what you've already done and what we've already liked, and write me a new post. And that seems to be working pretty well. So it's, you know, it's still not 100% by any stretch, and sometimes it still puts in the 10 million emojis and I've got it to stop single. So.

Deanna Shimota 00:12:45  When you create content or you get responses from ChatGPT, it's got the up hand signal on the downs hand signal. Do you give your GPT feedback?

Greg Padley 00:12:58  Sometimes I do. Most of the time I don't because usually I have something else to say after it writes something for me, unless it's close enough. And then I just say thank you.

Deanna Shimota 00:13:09  Okay, okay. I was just wondering if the experience is different. If you start to give it the feedback that it knows. Okay, I thought your response here was crap, but this one's pretty good.

Greg Padley 00:13:19  There's also that sometimes it comes up where it'll show you like, do you like this answer better or that answer better? I've gotten that a bunch too.

Greg Padley 00:13:27  I think ultimately it does help. I just I feel like if I just put a down thumb like that doesn't really give it feedback other than it got a down. Some like, how does that help it get better? I don't know how that would write.

Deanna Shimota 00:13:41  So one of the things that I've been using ChatGPT for that. I found it actually does a really good job of is cleaning up transcripts from videos and podcasts, and also identifying spots in podcast episodes that need editing. So this podcast, the Demand Gen fix, for example. And I've also got the HR Tech Spotlight podcast. I always pull a transcript of each of those, and what I'll do is upload the transcript and say, clean this up. So it reads well, because we like to use our transcripts on the actual pages that we post this on for the SEO value. And because not everybody wants to watch a video or listen, they'd rather read about it. And quite frankly, what I've learned with transcripts in these conversations is the way that people talk does not always read well, and it's totally comprehensible when you're in a conversation with someone in your talk, but when you're reading, it's like this is a little bumpy.

Deanna Shimota 00:14:46  They pause. They throw in similar words. There's all these things in the context of having a conversation that you wouldn't because it's not so simple and clean. So that really helps. And I found also from a video standpoint, when I would upload the video and I'd say identify the spots in the podcast video where editing is needed, don't edit out for content, but edit out for, repetitive words, grammatical issues, things like that, and identify the time mark for it. When I did that, I got a nice list back that was like, here's the six things during the conversation that should probably be edited out. I'm like, perfect. Now, to be fair, I'm not doing that with the raw videos, putting it in there and having it. I'm doing it after it comes back from the production team and rather than me listening to the conversation for half an hour because I was already part of the conversation. I just run it through ChatGPT to ask them to identify any spots that our producer might have missed, which tends to be pretty minimal, you know, with our editors.

Deanna Shimota 00:15:59  So I haven't tested it on raw videos and don't actually know how it would perform in those situations. Right.

Greg Padley 00:16:06  That's similar to having it do proofreading for you or give you writing recommendations. Right. It's the same thing, but you're just giving it a different you're giving it a different format to start with with this. Pretty cool. When you when I use it for proofreading and writing recommendations, it does the same thing. I'll ask it for whatever errors it finds, and I ask for provide recommendations and also ask it to provide a rationale, because sometimes we don't want to change a sentence and that'll but then it'll tell me why it's changing the sentence. So it'll give me a list of all of those things that it wants to change. And then the end it'll say, and here's the document or here's the the copy with all of the changes made in it so you can go through and pick out. Okay, I like the way I say that because that's the way I write. I'm going to leave that, but I appreciate that it's pulling it out.

Greg Padley 00:16:53  But certain things you leave and certain things like, oh yeah, that that works a lot better. I like what I like with GPT said about that better than the way I wrote it.

Deanna Shimota 00:17:01  Right. And I think proofreading is a great use of ChatGPT. I've had really good luck with it too. When you upload something and you're like, find the issues misspellings, grammatical spacing and identify exactly where in the document that those things are. It comes back with a solid lift and you're like, sweet. I didn't catch that when I proved it myself, but the machine did for you.

Greg Padley 00:17:24  Yeah. And sometimes it's like silly little things. How would you even notice? Sometimes it's just like an extra space or something or an extra period, and it's so tiny on your screen that you don't even notice that it's there. And you're like, oh my gosh, where did that come from exactly?

Deanna Shimota 00:17:39  Another, thing that I have found that ChatGPT is incredibly helpful for our summarizing meeting notes or reports and pulling insights from, like, interview transcripts.

Deanna Shimota 00:17:55  So, for example, I have some clients that I do consulting sessions with where I am coaching them on how to build their demand generation engine, and they're going back and doing all the work. So it's an hour and a half focused conversation where we're talking about all these things, and instead of me having to create notes or them having to create notes to document all of that, we can take the transcript from the conversation. Because I have an AI tool that comes to meetings with me and records that and, and creates a transcript. Right. But that I can then take that transcript and feed it to ChatGPT and start to say, identify the key themes from this conversation. Identify the to do's. Identify the important to note things like I give it the information of how I want to sort it out, and I say, create a bulleted list in these categories and it will pull that information. Now, before I hand it over to the client, I go through and I clean it up. So it makes a little more sense.

Deanna Shimota 00:19:03  But I would say in most cases it probably gets me 80% to where those notes need to be to pull that all together. So I've found that's really can be a time saving and useful tool when you have things like transcripts and reports and long articles and things that you need to summarize.

Greg Padley 00:19:22  Yeah, yeah. That's great. I like what you were saying about create this for me because I end up doing that a lot. Create a timeline and task lists for Dot, right. Create a table from this. Create a bulleted list. And so you don't have to sit there and type that and hit the return button and click on the add the bullet in and all that junk. You can put it in a table. It'll put it in a table automatically for you. Give it the data and say put this in a table. Column one should be this name. Column two should be that name. And it takes her info and puts it into a table for you. Instead of you cutting and pasting into all the little cells, or doing something silly like that to try and make it work.

Greg Padley 00:20:01  Just time saver.

Deanna Shimota 00:20:04  Right? And you have to be able to give it enough information and clear prompts to get what you want. Because I have had the experience where I've tried to pull things into a table and it just no matter how many ways I phrased it, I felt like I couldn't get where it needed to be. That's the joy of working with AI is sometimes you're amazed at how well it does something, and sometimes you're so underwhelmed that you're like, I just wasted half an hour trying to save 15 minutes. Yep.

Greg Padley 00:20:34  Yep, that's definitely one of the challenges with working with AI and really any new software, if you think about it. Right. So it all takes its big learning curve. So it's no different with AI. It's just we need to invest a time in it and see how it works and try to understand it. And it's not right. It's not going to go away. So you got to learn how to use it, and you're going to have to use it in the future, because it's just going to be the way things are done.

Deanna Shimota 00:21:00  Right. Another example where ChatGPT surprised me, how well it worked. I came back from a conference and I had a bunch of written notes, and the team wanted to see the learnings that I got from the conference, and I was like, okay, it's going to take me a few days to get around to actually typing the notes up. And someone on the team said, why don't you take a picture of it and ask ChatGPT to create the notes? And I thought, my handwriting is pretty sloppy. I don't know that it's going to work, but I took a picture, I uploaded it and I'll be darned. But it came back with notes. I asked it to sort the notes and before I knew it again, I probably had 7,580% there, maybe the 90%. It was pretty good how it captured the notes and how it categorized them for me. And I'm like, wow, this just saved me actually quite a bit of time because one, it probably would have taken me two weeks to put aside the 45 minutes to type up all my notes from the conference and to, like I said, it probably would have taken me 45 minutes to type up the notes, and in this case, it was a 15 20 minute project.

Deanna Shimota 00:22:14  Because I uploaded it, I gave the information, and then I just had to go and clean up and refine my notes from there. So it was quick and easy and saved me over 50% of time.

Greg Padley 00:22:26  It's funny you say conference, because it made me think of when you're at an event like that or a presentation of some kind, and there's slides up on the screen We see people taking pictures. So you can take all those slide pictures that you have on your phone that you never look at again, and put them into ChatGPT and get the content copy out of them. Oh yeah. And you can start start creating a creating your own database. Right. Things that you want to remember or things you want to talk about or things you want to share. Oh, I really you take a picture of the slide because you think there's something valuable there. But then once you take the picture, it's when you write your notes in the notebook and you never do anything with them. Once you've gotten the copy out of it with a GPT, you can can you have it and you can start using it for stuff and that are just sitting in your phone.

Deanna Shimota 00:23:12  Right. Yeah. Okay. So there's obviously more things that we've tried with it. We've got the list in front of us and we're already we've talked for several minutes here, so we're not going to have time to go through everything. But hopefully this gives those of you listening ideas on how to think about where you can use, whether it's ChatGPT or some other AI tool similar to that to help streamline and create more efficient processes. And I think the key is to step back and think about what are things that I do that are maybe inefficient or that are very time consuming, that I could potentially shave some time off. And the first time you try to do some of these things, it may take more time to figure it out. I can tell you, once you figure out some of those things, like it saves you so much time moving forward. I used to spend a ton of time listening to the podcast, manually writing out where the problems were that needed to be edited out, for example.

Deanna Shimota 00:24:15  And now it's a ten minute exercise. So there are opportunities, but you have to test and learn and figure it out to make it work. And once you do, it can be a really valuable tool, actually for us marketers to to help streamline some of the more administrative things that we do and even the spark some creativity.

Greg Padley 00:24:34  So remember, AI is here to complement your thinking and your work. Start small. Play around with it. Experiment and use it more and more than you would create. Once you start using it, you start finding out other ways to use it. It's like the first time you have a piece of chocolate and then you go, oh, I like this. Okay, I want a chocolate cake, I want a chocolate cookie, I want a etc., etc. so it's the same thing and make sure to tune in to our next episode. We'll talk through some more things that are working for us, and some things that aren't working for us, and learn some more goodies.

Jenni Geiser 00:25:08  Thanks for joining us on The Demand Gen Fix, a podcast for HR tech marketers brought to you by GrowthMode Marketing. Hope you enjoyed it. Don't forget to subscribe for more perspectives on demand generation and B2B marketing strategies. Plus, give us a like and tell your friends. We'll see you next time.