S2 E4: Marketing Artificial Intelligence with Brad Kostka

Adrian McIntyre: 00:02

From the Public Relations Global Network, this is PRGN Presents. I'm Adrian McIntyre.

Abbie Fink: 00:07

And I'm Abbie Fink, vice president/general manager of HMA Public Relations in Phoenix, Arizona and a founding member of PRGN. With public relations leaders embedded into the fabric of the communities we serve, clients hire our agencies for the local knowledge, expertise, and connections in markets spanning six continents across the world.

Adrian McIntyre: 00:28

Our guests on this biweekly podcast series are all members of the Public Relations Global Network. They discuss such topics as the importance of sustainability and Environmental, Social, and Governance programs, crisis communications, content marketing, reputation management, and outside of the box thinking for growing your business.

Abbie Fink: 00:48

For more information about PRGN and our members, please visit prgn.com. And now, let's meet our guest for this episode.

Brad Kostka: 00:59

Hi, this is Brad Kostka. I'm President of Roop & Co. We're a strategic communications firm located in Cleveland, Ohio. We primarily serve B2B manufacturers and professional service firms, particularly those in the financial services realm.

Abbie Fink: 01:16

Developing strategic communications over the years has certainly changed and evolved. One of the things we talk about is access to information and technology has certainly helped us change the way we access information, and along with that comes a real responsibility about how we're using that information. This idea of artificial intelligence marketing or using artificial intelligence for marketing is really something that all communications professionals should be thinking about, but there's a responsibility in using it and how to use it correctly. Something I know you've been diving into and using and wanted to chat a little bit about, really, what does it mean to use artificial intelligence and use it in a responsible, effective way?

Brad Kostka: 02:06

Yeah, that's a great question. I'll date myself, going back to the early days of my career over 25 years ago, some of the technologies I had to use from faxing out news releases and loading physical slides into a carousel for a slide projector. So the technology has certainly advanced quite a bit from those early days. Artificial Intelligence is something that's emerging right now, and I think we as marketers and PR professionals and communicators really need to be aware and on top of the technologies that are coming down the line. I think it's going to be as transformational as the internet. Dating myself, I can remember setting up our first email address. It was a company address that was thejamesjroopcompany, all spelled out, @ MSN .com was our very first email address. We used Microsoft Network was the provider of that email service, and it was a shared company email address, which is kind of funny to think about, and even setting up our first website. Technology has changed, and we as communicators have changed and adapted with it. I think this is another one of those seminal moments or technologies that we really have to be prepared for and get smarter about.

Abbie Fink: 03:19

Well, I can remember CompuServe with that "bum, bom, bum" when you plugged in ...

Brad Kostka: 03:22

Yes.

Abbie Fink: 03:24

But I can remember when we got email at the office and we would send an email to the person in the office next door and then run over to see if they got it, because we just couldn't quite understand how I could send something and it would end up in your office just seconds after I hit the send button. But, what about this is different? It's about using the technology, but it's really about diving into the behaviors and what the algorithms tell us and how we can help make decisions because of actions that are being taken or information that our clients and customers and their customers are using. Different than maybe, although we can glean a lot from actions taken in email and click throughs and all the other things, but it's really about looking at it from a data perspective and what we can do with that information.

Brad Kostka: 04:17

It's really about making us smarter as communicators and making us more effective. The campaigns and the programs that we're running on behalf of our organizations or our clients, it's about making them better and more efficient. It might be helpful to set the stage a little bit with defining what marketing artificial intelligence is. Let me read this to you, and I think you'll get a kick out of this. I'll clue you in here in a second. This definition I came across is, "Marketing artificial intelligence refers to the use of advanced technologies such as machine learning, natural language processing, and computer vision to automate and improve various marketing tasks and processes." What's interesting is that I used an artificial intelligence tool to actually write that definition. There's a number of tools out there. One that's pretty hot right now is Chat GPT. It's getting a lot of buzz right now. I plugged in, write me a definition for marketing artificial intelligence. That's what it gave me. That's not edited or anything by me other than me prompting it. I think that helps set the stage a little bit about what these technologies are already capable of doing.

Abbie Fink: 05:27

They're capable of doing a lot of different things, creating definitions, writing content, but there's got to be a human element to all of this. I mean, it's being written for consumption by, one would expect, a human of some kind. You can't just rely completely on what is being generated. That's a starting point in my view. It's got to have some other analysis and depth to it in order for it to actually make it be effective.

Brad Kostka: 05:59

Oh yeah, sure. There always has to be a human element to it. These things will get better, but for now, they're in the early stages and there are limits to them. They're also only as good as the human, I think, that's driving them, in terms of asking the question or in terms of the algorithm behind them. Then we could get into some of the problems or ethics, or whatever, that come with artificial intelligence, because there are some. I mean, it's not a perfect solution or silver bullet to every marketer's problem or challenge. But they are tools that we just need to look at to help make us better and smarter and do a better job at what we do.

Abbie Fink: 06:37

Where does it come into the process? When do you start discussing that we want to embark on this effort and we're going to use AI, artificial intelligence, to help us create campaigns? Where in the development cycle of your communication strategy do we start looking at implementing this?

Brad Kostka: 06:59

There are a lot of tools out there right now that have artificial intelligence behind them. It's everything from creating content, like written content. There's AI that's actually creating imagery right now, which is crazy. We've tried both of these. We are leveraging some of them for our clients already. But there are also some that can help with email management in terms of distribution or managing lists. They can help with social media management in terms of when to send messages out based on the data, when your messages would be most effective. They can be used for chatbots on a website to create automated responses where it looks like it's actually responses coming from humans, and those are improving. We have a client that's using that, and people start asking them, some of them have asked personal questions or are you real and things like that, because it actually can't tell the difference. I think it's really about looking at the different tactics and campaigns and things that you're running as a communicator, and identifying maybe where you have some roadblocks or some challenges or looking for you want to improve the processes or what you're doing. Then seeing, exploring what tools and technologies are available to help you in that particular realm. I mean, there's so many of them you can't possibly be familiar with or adopt them all, but I think it's good to start with maybe a small project and something that you can have some success and a home run with. If you get success there, then you can build upon that for other aspects of what you're doing.

Abbie Fink: 08:32

I want to dive in a little bit on one of the points you made about that you started to use it to create some content and to create some creative graphics and such or imagery. What struck me when you were saying that is, if some of this is me telling the technology to do something, if I tell it to do something and use the same instruction that you tell it to do, do I get the same piece of content back? If so, where does original work, copyright protection, intellectual property come in when if we're all feeding in the same thing and it's spouting back out the same thing, how do we make sure it's unique to our own brand?

Brad Kostka: 09:18

Well, I think that's where, if it is a true artificial intelligence tool, it's creating original content. It's not just like a search engine where you type in a query, and it gives the same results to me that you would get and a list of results. It's also not just, say, copying and pasting from multiple websites and in the same manner. It's actually, at least when it comes to the content creation, some of the tools we've used, it is using the data that's available out there, but it's then synthesizing that data and creating new text and new content based on what's out there and available. The same comes with we tried some of the imagery tools, and you can type in a phrase of, give me a picture of, we used one for a client that does security and we typed in some phrases about, show me a picture of an apartment complex with a gate and some security cameras, and it produced a number of different images that we were able to use for a website project. They were not images that were just grabbed from the internet, they were original images that were produced. There was no such apartment complex with that same gate and that same camera. They were originally generated. It is original, and so it is different. When it comes to the content, we're certainly not just taking and copying and pasting that. We're vetting the information and generally revising it, but we're starting to use it for things like you can plug in some audio and ask it to generate notes from a conference call, or actually we're using meetings and we'll create transcripts. I'm sure a lot of people are doing that now for meetings. That's AI that's listening to the audio, but then you can even take that transcript and throw it into another AI tool and create the notes from a meeting. There's a lot of pretty cool ways that these things can be used to make us more efficient.

Adrian McIntyre: 11:07

One example on the media creation side that I was just playing with recently is, I took the transcript of a long interview, fed it into a tool that creates a summary. Now, I took that summary, it was about 200, 300 words, and fed it into another tool. I said rewrite this summary in this format, and I gave it some specific guidelines. It actually rewrote it. Kept the ideas in place but rewrote the thing. Then I said, now take that new summary that you just wrote and write me 10 headlines of 50 to 80 words each from that summary that are speaking to professional services firms. It wrote out headlines of the summary, and then I have to select them. I have to work with them. I have to figure out, well, this one's not exactly right, I'm not going to do that. But it strikes me that this is an innovation in the way that we haven't seen for a long time in the tools for creative use. Nobody's complaining about the fact that every pencil, almost every pencil is the same shape. It's yellow and sharp in one end and sometimes has an eraser on the other end, because who's using the pencil is going to be different. I mean, language itself is already out there. It's just Hemingway used it one way and Yeats used it another way, and it's the same language. How do you sell this to clients? How do you say to them, "Well, look, we're actually going to have some machines do the work"?

Brad Kostka: 12:29

We're just kind of in the early stages of exploring it and using some of it. But I think on one hand, with some folks, that can certainly be a selling point that, look, we're ahead of the curve and we're using some of these advanced technologies and techniques, and it's making us quicker and more efficient and allowing us to spend more time to be more strategic, where we're automating some grunt work tasks. That's actually a good thing. It comes up in conversation, but I wouldn't say we're not that far along with it yet that we're saying we're an AI-driven agency. We're not marketing ourselves that way, per se, but we are using some of the tools to be smarter and better.

Abbie Fink: 13:09

Well, as you were talking, I was thinking about when we would hire professional photographers to come do head shots or check presentation events, or the grip and grin as we would call it. As more of us got comfortable with a digital camera, and then obviously when our cell phones became such an extension of who we are, taking those photos ourselves or shooting our video became sort of the easy way to make that done. It didn't replace professional photographers. There were still opportunities and needs that needed to be fulfilled by someone with the right tools and tricks to the trade that we weren't going to be able to do with our cell phones. I see this in much the same way. These are tools that we use to be better at what we're doing, but it can't replace a good strategic communications professional on your team. If you use it for that purpose, then perhaps you're not ready for the larger investment into a communications team. What I'm wondering about is thinking also around, what skills do we as professionals need to have in order to access this information correctly? Research and certainly data analysis and those things were not what I learned in journalism school. I suspect, however, that students that are graduating from the journalism programs, the public relations programs now, this is probably part of their curriculum and they're probably learning about the ethical implications of using it because they're going to write their theses if they're not careful by using this, so there's got to be a lot of discussion. But, what do we need to know? How do we need to understand this so that we are making intelligent decisions about when to use it and in what formats to use and advise on?

Brad Kostka: 15:16

Yeah, I mean I think it comes with this. Anybody listening to this podcast is educating themselves a little bit more about artificial intelligence, and I think that's really the starting point as to how we as marketers and communicators can do that, is just I think it starts with education. There's actually a professional organization called the Marketing Artificial Intelligence Institute. They started up a number of years ago and they have an annual conference, and you can get daily or weekly emails from them, and they have various trainings and things. That's a good place to start in terms of getting educated. I think it's also a matter of trying some of these tools and techniques. Yeah, I certainly got into communications probably and went through journalism. I got a journalism degree, much like you, Abbie, because I didn't want to be dealing with numbers or crunching numbers and things like that, or it wasn't a strong suit or whatever, although I'm better at it now then. I think it's about enhancing your team, then, with somebody that does have that skillset set, because somebody that's got that skillset doesn't have maybe the creative or communication skills and strategies that we have. I think it's just enhancing your team with somebody with those kinds of skills is going to be an important next step, for sure.

Abbie Fink: 16:37

What do we do as it relates to privacy and protecting that data? I mean, if you log onto a website, now it's telling you that the cookies are activated, or you're getting prompts to let you know that you voluntarily came here, but we're going to capture this information. You can't purchase something online right now without being fed, you might if you bought that, options. There's some responsibility to those that are mining for that data and using that data to be responsible about it. Then ultimately, those of us that use it to go back outward with it. What's the privacy implications? Where does the consumers assume their privacy, and where do us that use it from a marketing perspective need to pay attention?

Brad Kostka: 17:28

Certainly the use of AI, there can be privacy concerns. But, I think that ties back into just your general privacy policies as it relates to the collection of data and that you're properly informing people of that and that you have the proper protections in place. I think AI is just an extension of that privacy policy, and maybe you need to have some sort of additional language or something like that if you do start using some AI tools. But, usually your, I think, privacy policies and things do cover if you're using third party tools. A lot of us use a third-party email tool. Nobody has their own custom-built email distribution tool, and so I think the rules would apply for artificial intelligence tools as well. I think maybe where there's a similar area for concern could be the ethical challenges or the biases that might come in there. It could be a good idea to maybe have an artificial intelligence ethics policy in terms of how you might be using that to target customers. A quick example would be pretty interesting that I had run across as it relates to biases that we have to be on the lookout for. This particular case I ran across was with Apple. Apple had, I believe, a credit card that they were promoting. There was a husband and wife who both applied for this credit card. They both had pretty similar careers and salary levels and whatnot. When they applied for these credit cards and credit histories, all that was very similar, but when the results came back, the husband had a much higher credit limit than the wife did. There was some gender bias in the tool that was assigning the credit limits for the credit card. They actually raised this online and there was a bit of some bad PR that Apple got from it. But Apple then went back and looked at the algorithm, and I'm sure it wasn't anything that was intentional, but it was just that there was some general bias in this tool, and so they went and fixed the tool. I think that's a pretty illustrative example of something that we have to be concerned about when it comes to artificial intelligence and that we do have to look out for bias that might be in the tool, and certainly protecting the privacy and how the tools and things are being used, for sure, and the data is being used.

Adrian McIntyre: 19:47

Thanks for listening to this episode of PRGN Presents, brought to you by the Public Relations Global Network.

Abbie Fink: 19:53

We publish new episodes every other week, so follow PRGN Presents in your favorite podcast app. Episodes are also available on our website—along with more information about PRGN and our members—at prgn.com.