Localizing Global Marketing Through AI: How to Utilize New Tools in B2B Manufacturing MarketingLocalizing Global Marketing Through AI: How to Utilize New Tools in B2B Manufacturing Marketing

Localizing Global Marketing Through AI: How to Utilize New Tools in B2B Manufacturing Marketing Transcript:

Jeff White: Welcome to the Kula Ring, a podcast for manufacturing marketers brought to you by Kula Partners. My name is Jeff White, and joining me today is Carman Pirie. Carman, how are you doing, sir? 

Carman Pirie: All is well, all is well, and good to be here joining you for another episode of the Kula Ring. 

Jeff White: Yeah. And our topic today is one that I’ve wanted to be able to cover for a long time, but we haven’t really had an opportunity to have a guest with in-depth knowledge. About it, not to the level that I think the person joining us today has. 

Carman Pirie: Yeah, it’s a funny thing. I think we’ve some, yeah, folks, we’re talking about AI and, we’ve had folks that have wanted to come onto this show and talk about how they’re leveraging AI in their work, and then, frankly, often it seems to be a bit of a mile wide and an inch deep and… 

Jeff White: And mostly about content creation 

Carman Pirie: Yeah.

Jeff White: And expanding that side of things, and not really taking advantage of the intelligence part.

Carman Pirie: And then I think there’s almost the opposite side is sometimes true as well. Whether you’re following, getting regular emails from Gartner or some other more think tank bleeding edge AI usage that is out there. And I think sometimes those can seem a little cart ahead of the horse, if you will, for small industrial marketing teams that are really looking to just start using it and get it delivering results for them.

So I really think that today’s guest is meeting our listeners where they’re at with any luck at all. And so yeah, excited for today’s show. 

Jeff White: Me as well. So joining us today is Karlaa Gregory. Karlaa is the digital marketing manager at Sumitomo Drive Technologies. Welcome to The Kula Ring, Karlaa.

Karlaa Gregory: Thank you. 

Carman Pirie: Karlaa, I want to know about you and I want to know about Sumitomo. Where do we start? Let’s start with the company first, maybe introduce us to the firm if you would. 

Karlaa Gregory: All right, so Sumitomo Drive Technologies is an industrial manufacturer of gearboxes. So gearboxes make everything go from conveyors to the things in the airport that keep you moving, like the moving walkways, the jet bridges. You can find this in almost every single thing that you see. In your day to day, believe it or not, even roller coasters. 

Jeff White: I always find people don’t move enough on those ones in the airport, but that’s the topic of the time.

Karlaa Gregory: Yeah, they do move a little slow, but other than industrial gearboxes, we do have other product ranges for motion control products like pharmaceuticals or robot arms. That’s the best way to explain it. Variable frequency drives, repair. We do a lot and we are a global company. 

Carman Pirie: Very cool. Very cool. And where are you based?

Karlaa Gregory: I’m based in the U. S. headquarters, which is in Chesapeake, Virginia. But we have offices in Canada, Latin America, Brazil, and of course other continents as well.

Carman Pirie: Well, Karlaa, it is wonderful to have you on the show. Perhaps we can learn a little bit more about how you ended up at Sumitomo. Tell us a bit about your background. 

Karlaa Gregory: So I went to school for communications, and marketing communications at a local university, and I started off working at Liberty Tax. So Liberty Tax is one of the bigger tax firms, not as prominent as Jackson Hewitt or H&R Block, but they were up there. They probably still are. I did social media for them, for the whole company. And then as they were moving to Dallas, I decided not to go and I moved over to Sumitomo. So I started there in 2019 as a digital marketing analyst. And digital marketing was never really at the forefront of the strategy like there were little things that we would try to do or that the woman who previously had the position would do. And I took it and made my own. So doing more SEO, doing more social media management. And eventually, I was promoted to digital marketing manager. So now I lead a team and I help the other regions as well. 

Jeff White: I love the immediate jump from a not old school, but an old school type service sort of business to a high-tech manufacturer and many more digital initiatives all at the same time. Yeah, very interesting.

Carman Pirie: Well, Karlaa I know that you and the team have been working on making some kind of practical implementations and use of AI in your work. My understanding is that spans everything from more like persona bots, type of, or persona CBTs to help you understand who you’re marketing to or even understand the products that you sell at a deeper level, through to reporting. So I think we have all kinds of stuff to cover here. I’m curious, where did you, where do you start dipping your toe into it? What was the first initiative that y’all took? 

Karlaa Gregory: The first thing I did when AI started to become bigger and bigger was make my own GPT to help with the other regions. So I uploaded a lot of our catalogues, brochures, and any documentation I could find. Also, I made sure to upgrade our chat GPT account to the team’s account because we don’t want that proprietary data training the model. So I made, I made that, I added some prompts, I’m adding some actions and just some guidelines. So I can go in there and I can ask for assistance with a campaign or what’s the best way to start this campaign because since we are a global company and I do try to help the other regions, not everything is the same.

Latin America might have a different cultural way of saying things versus Brazil versus Canada versus the US. So I definitely use that there to get me started. And then of course I do run it past the team leads, the marketing leads. In each region. 

Carman Pirie: Help me understand that. So you’re using basically the uploaded all the localized regional specific catalogues, literature, et cetera. You’ve trained the model on that and now you’re using that to help you generate localized campaigns. Do I have that correct? And that’s fascinating. So I’m curious, how are you running this back by the team leads and saying, are we getting it? How close is this? So that’s my question. How close was it? And I’m assuming it’s getting better. 

Karlaa Gregory: Yeah. So it’s always getting better. At first, it was a little off, but. It’s getting better. It has the personas correctly. Some of the content is the same. I haven’t used it so much to translate because they do speak Spanish or Portuguese. So sometimes I’d rather that be natively translated just because there are so many different nuances and, how to say something like gear motors. When I use Google Translate, it says it’s pronounced this way, but actually it’s this way. And then. I learned the hard way that sometimes Argentinian Spanish is not the same as Mexican Spanish. So I let them handle that there, but just the overall concept of the campaign and how to go about it has definitely helped with this GPT. 

Jeff White: What was the initial impetus to decide to explore AI, was it increasing your capacity as a team or was it to find out some more specific ways of communicating in the different regions? Like how was your mindset as you were going into this? 

Karlaa Gregory: So what happened was when I got my new job title, I missed a nice little portion at the bottom that said to assist the other regions with digital marketing. So my first year as a manager, I did not do that. And I realized that I don’t have enough time, I have enough time, but to try to help and lead the way. Can be very time-consuming. I love the team and I love helping them, but still have to manage the U S and then also duplicate those efforts for three other regions. And so I was like this AI thing is cool. You can make your own. How can I do this? So I just decided to sit down one day because at my job, we do get learning hours or I know I build that in for at least weekly or maybe monthly just to sit down and learn or just do something, read a book, listen to a podcast like this one, or, just anything else. So I sat down and I just started chugging away at this GPT. 

Carman Pirie: All right. Let’s dive into where else AI is coming to life because I know that you’ve done more than just this GPT for your regional work. Where else have you seen the remaining initiatives, what would you say you’ve seen the kind of the biggest impact? 

Karlaa Gregory: I think the biggest impact is the ones that I have made for different personas. I’m still working on them, still trying to make them better, but two of our main personas, when we are looking to acquire leads and just, do more brand awareness, are engineers and procurement staff, because that’s someone’s gonna draw out or figure out what, how they want the gearboxes fit into their operation and then procurement buys it. And a lot of times it’s really hard to maybe, interview one of our engineers in the office or just try to figure out what they like. So, I asked some questions to our engineers. And then from there, I base that on the GPT I made. So I have one that is an engineer, and I have one that is a procurement specialist, and I can just go in there, I can ask questions, I can say, what’s the best content type that would resonate with you? Are you looking for more technical data versus more non-fluff data, what’s your preference? So using that to have my own engineer or procurement person has been really helpful. 

Jeff White: What surprised you about what your engineer or persona or a procurement person has said to you as a response?

Karlaa Gregory: Let’s see, there was one where I was confused and I got really mad at it. And it just was like, I’m listening. Like it, it almost turned off. I swear to you not, it was upset. I think just even the data-wise, like looking at the two different ways to address the same type of content. The procurement people, really care about cost. While the engineers want really good tech specs, like they want to know everything from torque to all the major specs that you need to select a gearbox versus procurement wants to know what’s the value, how long is this going to last me, is this going to, is this gearbox going to break within two years or, can I Opex this or Capex this for 10 years? What’s the difference? 

Carman Pirie: Yeah, I think I suppose a lot of people, you instinctively know that if you’re speaking to procurement or you’re speaking to engineers, you’re probably going to have to talk to them differently. But I suppose a lot of marketers would struggle with how to do that or how to bring that information to life for those two different audiences.

So are you using this for content that you know, is specifically going to be channeled to those personas directly? I’m just trying to understand how the impact of this is coming to life. 

Karlaa Gregory: Oh, yeah. So I’m using it to formulate any campaigns, so any ABM campaigns that I’m doing, including LinkedIn ads. So some of them, you can do career targeting or job title targeting on LinkedIn. So I use that there especially. 

Jeff White: Is it producing the imagery for you or primarily the kind of headlines and campaign concepts? 

Karlaa Gregory: Mostly just campaigns. Mostly just content and the campaign concepts. So we do have two graphic designers on staff, so I allow them to do that. But sometimes I will ask what type of imagery and kind of give her those guidelines to work with. 

Carman Pirie: Very cool. And have you seen, I’m assuming you’ve tracked the effectiveness of those campaigns before you were using this technique and after as well. Have you seen a difference in how they convert?

Karlaa Gregory: Yeah, I’ve definitely seen an increase in click-through rates, maybe not so much with lead gen, but lead gen from the landing page, so they’ll come around maybe later. So it depends on which area of the funnel we’re in. So for the top of the funnel and middle of the funnel, they’re doing great. Some of the campaigns or some of the persona bots that I’ve been using have generated middle-of-funnel content like maintenance guides or how to make sure that you’re, how to choose between CapEx and OpEx. So those are all doing really well, but we’re still retargeting them with other content to come back around. 

Carman Pirie: And you’re also using AI for reporting as well, yeah? 

Karlaa Gregory: Yes that’s my new baby. I’m really excited about that. 

Carman Pirie: Alright, take us through it. What are you doing? 

Karlaa Gregory: There have been times when I might be out of town, I might just not, my manager might be in a meeting and ask me a question like, hey, can you give me x, y, z stats really quick?

And I’m like, okay, usually it’ll take me like an hour or so to dig in that data, go to Google Analytics 4, do the explore report, try to find exactly what she’s looking for. But now I do use a mix of Zapier and API calls. So Zapier pushes the data to a Google Sheet, and the API calls go from the Google Sheet to ChatGPT.

So the data is automatically transferred every month, so I don’t have to really worry about going in, exporting the report, or reformatting it. It’s automatic, and that is from all of our channels, including website, organic, and paid social. I had to manually do some of our trade publications, so sometimes I have to do those manually, but that’s way less than doing all of those manually.

So, since the data automatically aggregates in that Google sheet. The ChatGPT can go in there and ask questions. I have a list of questions that people can ask. So if they’re not sure of how to phrase something for the GPT to understand, they can ask what’s the best way to access it, and it’ll send them a Google Doc to click on. But so far, it’s been very helpful with just pulling up easy reports, my monthly reports. I have connected all of our conversions, and our key events on Google Analytics. It is beneficial for all of our channels, so that way I can pull that data in, I can ask how the social campaign really does because we love impressions. We love reach but no, when they click the link from that LinkedIn post How do they respond, like what do they do? 

Jeff White: Have you had to? train the model in order to understand the difference in kind of the importance and level of KPIs like that to move from more vanity-type metrics, like impressions, through to the pipeline or revenue generated.

Karlaa Gregory: Most definitely. So it, it took a lot of, it’s still taking a lot of training to get to fully understand what’s funny with when you make a custom GPT is that you can have instructions, but the instructions can only use 8,000 characters, which sounds like a lot, but when you want to be very specific about what you’re asking, you’re going to need more. So what I do is just, upload my own instructions on a Google sheet so I can have unlimited characters. So that way, when I see things that might not, like when I ask a question and it doesn’t come out correctly, I can check the instructions to make sure it’s listed. And then if not, tweak it as necessary and then retrain it to see the outcome.

Jeff White: That’s it. Very cool. Yeah, it is. It is. I 

Carman Pirie: Yeah, it is. It is. I can imagine there are a lot of listeners right now thinking that, okay, this dog might hunt, right? There’s a there, there because it could be being asked a number of times through any given week or month to provide some data to whether it’s the product development teams or other marketing teams for other divisions or what have you. The idea of a reporting GPT is to be able to get this information into the hands of people who can action it faster. It obviously seems incredibly beneficial. I’m curious, Karlaa, how widespread in the organization is its use now? How many people have access to it? 

Karlaa Gregory: It’s just me and the marketing team right now. Just because when things aren’t 100 percent correct, I don’t feel comfortable really sharing it that much because then if something goes wrong or it’s not correct, it’s, oh, is this broken? I’m like, maybe you just didn’t ask it correctly, so until I get some better training with other people to use it, it’s staying in the department, but I also want to mention that it does visualize the data as well. So you can get the analysis of how the campaign is doing, but you can also ask for it in charts, bar graphs or line graphs. 

Jeff White: So at this point, you’re making sure that it still has a Human driver who will be able to sniff out if you’re maybe not getting the exact type of results that you should be. But eventually, it may be able to be released on the C suite so they can ask questions anytime they need to. That’s that is the goal.

I wonder, you know, you’re talking about using Zaps with Zapier to do API calls and feed things to Google Sheets and all your custom instructions and things that you’ve learned about using these tools. How, do I know that there would be a swath of marketers, that are listening to the podcast that would be hearing those things like, Oh my goodness, that’s just so technical that there’s no way that I would feel comfortable doing that. Do you feel like this is something you’ve learned as you’ve approached it? Or have you always had a technical mindset about how you approach this kind of thing? 

Karlaa Gregory: I’ve always had a technical mindset and I’m always a learner. So that’s one big thing at Sumitomo is that we always give our employees space to learn. And so I taught myself. I figured out API calls. And some out-of-the-box tools do it, but when you want, like you want to be really specific and you have things that you really need, it’s easier to make your own. Cause I’ve tried both.

I tried some app that allowed you to do it like that, but the way it worked. It was just so like clunky and cumbersome and you really had to know exactly what you wanted to ask. It didn’t really connect the GA convergence to what you were doing, and it was just but there, there are some out there, but I just prefer to make it custom because I want this to be a long-term thing and knowing that, Zapier is not going anywhere, but these tools might also de depreciate over time. So at least I have control over that data. 

Jeff White: Yeah, for sure. And it speaks to how, the big concern, of course, with AI is that a lot of marketing tasks can be replaced by it, but the fact of the matter is, what you’re talking about is an incredibly complex ecosystem of different tools and insight and knowledge to understand the output that you’re getting from them that, simply is not replaceable. At this time and probably anytime in the near future. 

Karlaa Gregory: Yeah. And I know I’ve seen people believe that I don’t, I honestly don’t think so. I think it’s going to make you better. It has definitely improved my output. And it’s also helped me learn. Anyone who works in the industrial manufacturing industry knows that it is difficult. My first six months at Sumitomo were spent trying to figure out the difference between a Paramax and a BBB4, which are two of our top products, or what a right-angle spiral bevel drill is. I could not figure it out. But recently we started to sell electrical motors. Industrial electrical motors, and I use chat to explain it to me, I actually ask it, explain this to me as if I were a fifth grader and explain, explain how the electrical motors work, how they integrate, because again, trying to find someone in person to talk to can be a little hard or cumbersome at times like they might not answer the phone, they’re busy doing their own thing, nobody wants to have to teach you, but sometimes people have better things to do, and that’s okay.

So using AI to train myself and learn, I’ve learned the most about industrial marketing using ChatGPT over the past year than I’ve learned in my previous three years at Sumitomo. And I always run it by him. So when I do make the content or I make the landing pages, I run it by all the top people in charge of our product managers. And very rarely have they said, this is a hundred percent wrong.

Jeff White: You got to have that initiative. 

Carman Pirie: Yeah, that’s it. I think you’re spot on there. There are a lot of folks, it’s not even just that the engineers are maybe busy or whatever, but sometimes a new marketer, especially, can be very intimidated about asking too many dumb questions or feeling like they’re asking too many questions. And you can get a pretty rapid head start by the clever use of ChatGPT and other like tools, I think. 

Karlaa Gregory: It’s definitely helpful. Something I did learn at a recent conference I was at, was that some people will upload transcripts from either they’re like C-suite or their product managers and then get that tone of voice and get that data into their own GPT. So then when these people, cause times are changing. That sounds terrible to say we all know the workforce is moving over, and people are retiring, but some of those great minds are going with them. Like they’re retiring. You don’t want to call Bob. Who’s been at Sumitomo since 1985, he’s on the island retiring. Like he doesn’t want to answer the phone. So if you can get some of that data or just get a transcript with you, just asking him questions, you can use that as well. And have your own virtual Bob. 

Jeff White: Wow, what an interesting idea to use this as a way of avoiding retirement and the absence of information. Holy cow.

Carman Pirie: It is interesting because we’ve certainly used AI to analyze qualitative responses, like interview transcripts, and things of that sort, to find patterns over a batch of interviews. Very helpful, and interesting to think about using that to actually inform tone of voice in future communications. That’s really interesting. 

Karlaa Gregory: Yeah, especially if you’re doing any ghostwriting for your CEO. It’s just hey, okay, this is what, here’s some things you say, this is how you speak. How can I use this to just write a post for you that sounds like you and not like a robot?

Jeff White: I think that’s really interesting. One of the things that I’ve realized while watching marketers who are deeply interested in AI is that they often really like to experiment with the new tools. We’ve talked a bit about ChatGPT and how you’ve used things like Zapier. But have you explored any of the other tools like Claude and all of those other ones? And is there anything that you’ve learned about the specific qualities of them? 

Karlaa Gregory: So I think I don’t know if the app is called Poe, but there is one app that has all four or plus, like all of them in one, and I’ve tried it, but I feel like I’m just an original chat GPT girly I would rather just stick with what I know. I know they do have different pros and cons, but GPT just seems to be the most robust. Especially with the different upgrades and how you can do API stuff. And I just, I’m just a fan. I have used AI to make different advertising creative. So my employee, she was on maternity leave and I was like, I need help. And there’s an app called, I think it’s creative AI maybe. And you can make one ad. And from that ad, you can make it into all the different display sizes. So instead of having someone sit there and make it, the long, wide one, the short, wide one, tall, you can just snap the finger and it will sometimes decently, sometimes it might need a little bit of some tweaking and stuff, but it saves a lot of time that way, especially when you want to get a campaign off the ground quickly.

So pairing that with the GPT that I’ve made, now I have the creative, I can at least, get the campaign on the ground, start getting some baseline results, just to even see if the targeting is on, and then from there grow, tweak it, optimize it as necessary. 

Carman Pirie: Karlaa, you mentioned that the reporting at GPT was your latest baby. So I want to look ahead, what’s that, what’s the next one that you’re going to use case that you think that you’re going to dive into? 

Karlaa Gregory: The next use case is the one that I just started, the conference that I talked about. Interviewing some of our retiring members of the staff and having a GPT there. Because again, we have people who know so much about Gearbox, and we’re all so used to just going up to Bob’s desk and asking him about something, again, I think he’s retiring next year. So we want to get all that data somewhere. It can be an audio recording, or it can be a video recording. It can be like white papers he’s written, anything that we could upload to beef this up. So then in the future, when we do need to talk or figure out, what’s the best way to explain this? We have that there. 

Carman Pirie: The Ask Bob project. 

Karlaa Gregory: I should call it that. 

Carman Pirie: I think that’s really cool. I look forward to seeing how that comes to life and basically using it for knowledge transfer.  Yeah, that’s going to be fun. This has been a fun show. I’ve enjoyed chatting about your use of AI and how you think about it. It’s been a joy to have you on the show. Thank you. 

Karlaa Gregory: Thanks for having me. 

Jeff White: Really inspiring. Thanks so much.