1123 Presents

4. AI - How it Started and How It's Going w/ Shana Boyd

• 1123 • Season 1 • Episode 4

Share episode

In this episode of 1123 Presents, CEO Shana Boyd joins us to discuss AI in the advertising industry. We start by introducing the approach we take with AI—not focusing on tools and how-tos, but rather sharing our real-world experiences and the impact AI has had on our business. We recall the early skepticism around AI and how, despite initial doubts, these tools have integrated deeply into our operations, enhancing both our creativity and efficiency.

We emphasize the importance of transparency with clients and internal protocols to ensure responsible AI use. The discussion highlights how AI assists in various tasks, from content creation to media planning, and how it complements human expertise rather than substituting it. We conclude by encouraging a balanced view of AI—embracing its benefits while maintaining a human touch in the creative process. 

Dave:

Okay, so welcome to another 1123 Presents. Today I have with me my business partner, co-founder of 1123, shaina Boyd. She's our chief executive officer, hi.

Shana:

Hello, thanks for having me.

Dave:

Good to see you again, and obviously I am Dave DiStefano, chief creative officer at the agency. We are going to discuss a little bit today about AI. We are not going to do what a lot of other people have done, which is tell you what to do and how to do it and what tools to use and all that stuff. We're just going to have a little chat and we're going to talk about what we have experienced in the last, I would say, couple years, especially in the last year of the advertising business, where AI has been way less of a fringe thing and more of a hey, everybody's using it for X, y or Z. Up to now, I mean, it has touched every, just about every aspect of our business, right?

Shana:

Yeah, I would say it is creeping in.

Dave:

Yeah, yeah, it's blasted through the door, kicked the door down so early on, when AI was kind of rolling out. There was a lot of things out there about like chat, gpt and people generating images and all that. Like I know me being the sort of like pragmatist skeptic that I am like it was like oh yeah, we'll see, you know, not really sure what. What were your feelings when stuff like that started rolling around and people started talking about it like just outside of our sphere?

Shana:

Yeah, so it reminded me a lot of my early days at agencies when, even like the internet, even though it was well around and it had been for a while, there was like a big play to start selling advertising on it. It was free at first, and so I don't know for that. It just reminded me of being like an early adopter to something and it was so new, you know. So I was like, well, I'm going to wait a minute. You know, kind of like I did even with that back then I didn't like jump all in to buy digital advertising right away until you could kind of see what it was going to do.

Dave:

Yeah, Are you an early adopter in other areas of life, though? I mean like phones, cars. Here's what's great. Everybody. I know the answer to this question.

Shana:

I'm just rolling that carpet right out there, the answer to this question.

Dave:

I'm just going to say are you serious?

Shana:

I'm just rolling that carpet right out there, right? No, I am not Right. I am not. I'm very cautious. I like to make predicted, calculated risks. Yeah, it's one of them.

Dave:

It's actually great for anybody who's thinking of hiring us. That's a great, that's a great trait to have, great trait to have in somebody who's in charge of accounts and media.

Shana:

Yeah, I'm in charge of your money. Calculator risks are always a safer play.

Dave:

Right. So, yeah, a lot of the misconceptions. There's a lot of fears. I mean, everybody was kind of going off into, hey, you know, this is going to be the Terminator, and Cyberdyne systems is going to take over and we're going to be, you know, fighting robots on a pile of skulls at some point. Um, it's not that. Uh, it's still.

Dave:

I think it still threatens parts of industries, but the way it's threatening those industries and those areas is it's threat. It's threatening people's way of thinking, and so it's become more of how people do business in a lot of ways. I know there's some areas of retail and things like that where you know, yeah, there there's labor that's changed, the labor landscape has changed and all that. But in our industry, in advertising, it's become more of like, I guess for me on the creative side of things, when you think about how Photoshop or or or Adobe has added features named drop, um, but like, if, if something, if something develops, you know things progress, there's a new tool out there. You can fight it. You know you can. You can fight the convenience or the. You know this is not the way I I'm used to doing this. But you, if you don't, if you don't figure it out and learn. I mean, if you're, if you have sort of a learner's mentality with this stuff, you can figure out how it becomes a tool in the toolbox, right.

Shana:

Right. No, I agree, I think to me it's like having navigation added to a lot of cars. You know, it's not scary, it enhanced what was already there. I think it's up to the individual to, you know, still make themselves useful at work. The people that are probably a little bit more scared about AI and kind of like not leaning into it, are probably more, I would say, like people who want to check a box or be like just very task oriented. I think people who are curious and still driven. You know they're utilizing the tools to enhance what they already do and enhance that expertise and not like letting the machines take over their job because they're they're more valuable at that point.

Dave:

Right, exactly, and we, I mean that's one thing that I think we've been very cognizant of as we've navigated this. You know, it's just the integration of it, not necessarily like it's, it's not one or the other. So I guess, what do you remember? What was the first thing? I think that we did, we, did we?

Dave:

I think it was Chad GPT, probably that we used first and we just sort of tested things out, like a lot of folks Like we. We definitely did some. There were some seminars and some education at the employees. Some of our folks listened to and and and attended either remotely or whatever, and there was, you know, there's a lot of education around it, but that all changes Like it's it's developing so fast. So we kind of dipped our toe in a little bit and I think we use chat for some stuff in the background.

Dave:

But we quickly learned like if, even if you start there for, say, like a thought generation thing, or if you're like, what happens if we say X to chat or what if? What happens if we want to mock X, y or Z up in Dalai or whatever, you don't end there. Even if you start there, you know there's a human there who's generating an idea. We want to achieve a goal, we want to try and create X, we want to try and get someplace. So let's kick off with this and use this as a thought starter. So we kind of morphed into that right.

Shana:

Yeah, yeah, no, I think. Yeah, chatgpt was one of the first things that we were testing. The media team was also testing the new AI integration tools. On Meta, I want to say that was like fall of 2023, that some of that was starting to roll out and that's you know where they. You basically dump in copy and different imagery or video clips and then you let the system automate and pick pieces of it. You know so that. And then, yeah, gpt, um, just throwing stuff in there and honestly seeing what it would output. You know um, and then, yeah, learning um some of the rules and regulations around it. I think, because, as we were testing, we were trying to figure out, okay, within these programs, like what is um, like what could be even copyrighted, you know what is public versus private? Um, how much do we want to let the machines take this?

Dave:

and run with it.

Shana:

Yeah, I know, I know I'm in terminatoring it, but how much do we want to let the automation, I guess, kick in versus kind of some of that human manipulation too? So yeah, I think finding chat GPT was awesome. I mean, it cuts down a lot of time on copywriting or even like research. You can kind of ingest a lot of data into that. But I think you know, what we found is you're spending more time on the back end, not, I guess I shouldn't say more time, but the time's more valuably spent, because then you're going through and kind of revising or reprompting or, you know, getting to that little like gem that you're looking for.

Dave:

Yeah, what do you think about, like content creation when it comes to that? I mean, we're touching on that a little bit, like, as far as content creation goes, you know yay or nay. I mean, where does it? Where does one stop and the other begin?

Shana:

Like um, I think a good uh case study around this. We tried to utilize it for a client and what we found is it just doesn't sound like their voice, you know, so if we were, if we were working with somebody and had been for quite a while and that brand has had some semblance of a voice online for content creation. In any way you cannot use GPT and let it output and just move along.

Dave:

Just flat. Yeah, you can't just be like, oh, it made a thing, so there it is.

Shana:

No, I think the thought starters or those ideations are probably key, yeah, but even with that, I want to be kind of careful because I feel like, oh gosh, if they're pushing this out to us, then they're probably doing this with a bunch of our competition too. So, like you still want to have that like unique creativity around content creation as well.

Shana:

So even if it thought starts like, and we use maybe three to four of the ideas, I still would maybe like and I mean you know this like we challenge our team to kind of come up with their own as well.

Dave:

Yeah, the voice thing. I mean the voice things come along, because once you, once we got through, uh, I think, over toT-4, whenever that rolled from 3 to 4, and now it's like, at least at the date of this podcast, it's 4-0. The way you can construct GPTs, you can actually put a brand's voice in it. But what I've noticed is when you're steering it, even when you do that, you've got to look and make sure that you don't get some sort of canned. You know there's, there's, there's always. To me there's always sort of a, a piece, like when you were talking about the voice. You can almost get there. It's very accurate, depending on what you prompt. Like it's only as good as you prompt or program. And then when you get back you're like, okay, let's clean this up, let's make sure this is, you know, set.

Dave:

So, I think again, like it's such a back and forth, it's such a. To me, like as somebody who was a skeptic before, now it's more of like okay, that's my whatever, that's my paintbrush, that's one of my paintbrushes.

Shana:

It's not all of them exactly, it's one of them um well, it's like media tools back in the day used to hand crunch reach and frequency and all the metrics and, oh my gosh, like you know you could spend hours doing that in your job. Well, we have the systems now where you just integrate the schedules and it prompts out those metrics. It doesn't make you less of a media buyer, it just gets you to move along and you can handle more buys. You know, I think.

Dave:

I think. I think, in a way, the there's still as a conception with a lot of clients, I think, on the client side, like oh, you did that in AI, or whatever. Oh, did you do that in AI, or I'm going to why don't you just do it this way? Why don't you put it through X, Y or Z?

Dave:

Well, we can, but you know, there's quite a bit of advertising experience in the brains of all the people in an agency, and so again, if they're, if they're using it just as a substitute for their own brain, you've got a problem Like. That's not going to work. Their brain needs to work in conjunction with the tools. My car is not going to drive itself down the road, even in cruise control, and be able to turn corners, you know. Maybe, it'll adjust its speed. I think cars have radar now and stuff right, but they don't turn. Like you're not going to get a car to turn, well they not?

Shana:

I don't know, actually Tesla's might. Are we sure about that?

Dave:

Can we?

Shana:

get a fact check.

Dave:

Your Tesla's might. Lots of name drops in this episode. Your Tesla's might but, they're definitely going to explode at some point.

Shana:

Oh, my gosh so there's that.

Dave:

No offense, elon, please don't come get me. What hurdles did we run into? Do you think that we needed to overcome at first, or what have we run into so far? You think that's like other than you know? I guess legal is one of the things you touched on.

Shana:

Yeah, Copyright things and all that which we haven't, I'd say legal, was a big one to me that I wasn't even really thinking about. And I went to a AAF sanctioned event and there was a lawyer that was focusing on AI and obviously for him it was still like evolving you know as you go, but the biggest thing that he brought forward was just the fact that if you create art within any of those programs, is it Dolly that, I guess?

Dave:

is one of them. Yeah, that's one of them. Yeah, so there's a bunch.

Shana:

But you know, I think he was maybe referencing Dolly at that point, but he was saying there's a lot of people that are spending hours and hours prompting and then when they went to try to put a copyright on it, to say it was, you know, unique to them, they couldn't, they don't get to own it. So that was eye opening to think, oh gosh, you know you can't use this for you know, some type of a national campaign or I guess, just like modern art, I mean even just you know, no matter what um, you have to go through a different step to get that approved. I I suppose there's ways that it could still maybe get copywritten if you go through some of those brand paid for subscriptions. But I actually don't know a hundred percent what. What's evolved and again, that's been almost a year, so I would imagine things have continued to change.

Dave:

One of the changes, too, is that the, the AI, that's open source, that isn't inside of any of those that can be owned, um, so that that is one thing that has changed from not too long ago to now. Um, so, yeah, I mean, I think the best thing to do in those situations and again, I'm not, we're not, we're not on here to get to do a primer for everybody to follow rules One two, three, four, but by the time we'd say it, there'll be antiquated, it'll be, it'll change.

Dave:

Yeah, yeah, your phone's obsolete when you buy it, but the big thing is to, yeah, if you have legal counsel in a business setting, for sure, and you're trying to do things, I would say you know, honestly, the biggest thing is to not let AI completely steer something you're creating, drafting whatever, without the human element getting involved and making sure this cross checks. This is the right voice, the right whatever.

Shana:

Yeah, but legal, and then also just well, and then on top of that, just not even taking it past the step of legal, but it's like making sure we tell our clients when we're using it.

Dave:

Transparency that was one of the things I was going to get to here is and we've got that at the agency as a rule is like, if we are going to use it for the generation of, maybe AB testing, for copy or whatever, like if we do any of that stuff and we're and again, we're still involved in the process, we're communicating like, hey, we're probably going to run this through X, y or Z To see what pops out, to see what we can get for variants. And hey, client, just letting you know that's a, you know that's, that's one of the things we're going to do. Hope that's okay.

Shana:

Yeah.

Dave:

Success stories. I mean things that we've done, I guess, with for us, like the way things that AI, the way AI has complimented our jobs. Now I mean for you. What do you think?

Shana:

Yeah. So definitely we use it, or I've used it, for writing up some different case studies, award submissions. I feel like it's actually helped maximize the efficiency and the amount of internal work that we can output. Maximize the efficiency and the amount of internal work that we can output Because, you know, I mean sometimes, like when we're advertising for ourselves, like that goes to the back of the list because obviously we're busy, we have client work to do. So I feel like it's been really nice for, like you know, help and support in that area. So we don't miss out on opportunities like that. Because I know for me, like just to submit awards, you know it would take me a while to kind of sit and think and really like write around it. Ai has definitely allowed me to put in like some high level achievements and then let it kind of help output, definitely using it with a lot of our digital tools as far as media placement goes. So we're letting you know the machines God, I keep saying that. I know that's a term.

Dave:

I really hate using that term I know we're using the machines. Look out, get down, you know. Oh my gosh.

Shana:

I know. So we're letting like some of the automation I guess take place. Like I'll tell you one of the newer things within YouTube, they have some AI, so once a month it will push out. Well, I guess, when we're doing our reporting, so on that once a month cadence, we're kind of taking a look at like that all around report and it'll actually give us some ideas for additional affinities or like targeting capabilities or like targeting applications that we should add to the campaign. And that's been nice, because it's sometimes it's things we don't really think of. Um, the AI plus integration with Facebook has allowed us to get an even lower cost per thousand. Um, they do tell you that. So I mean it's obviously their goal to, you know, get you to use it, so they incentivize you a little bit.

Dave:

Yeah, yeah, I mean creative, creative side. It's an interesting um. I find myself a lot of times like I'll hit GPT for like what's another word for this, what's another phrase for this? And kind of throw things around, kind of see what comes back on on some stuff we're working on conceptually, um, but I use it, we use it a lot conceptually Um, we, we, we really do. I would say that's the biggest amount of work we do in AI at the agency, in creative is conceptual.

Shana:

Yeah, that makes sense.

Dave:

I've been able to. You know, I sketch storyboards for some of our bigger productions. I love doing that, but in a lot of ways like if I put in a prompt and I know how to prompt, I can, you know, generate a couple of different boards and look at them and be like, oh, I'm going to sketch it that way. Oh, I'm going to look at that Like that's, that's a, that was a, that's a great use. That that I've used, that I've been able to do here with campaigns and again, it compliments the process.

Shana:

Yeah.

Dave:

People still got my drawings. You know our clients still got the creative director's vision for stuff. But it started with like hey, let me see this guy eating a hamburger and these four different angles oh, I like that one, I'm going to sketch this. I don't like that Because the other thing is, when you generate images, you don't get the same thing every time.

Shana:

Yeah, that makes sense.

Dave:

You're not going to get the same image every time. You can write the same prompt a little bit.

Shana:

Yeah Well, it probably figures. It's not giving you what you wanted the first time, so it's trying to like automate to you know, get closer to what you're looking for.

Dave:

But even day to day you could, you could pick up, you could do the exact same prompt in GPT for Thursday and then Friday come back in a brand new. You know, chat, start a new one, put the same prompt and you won't get the same exact thing. Yeah, colors, lighting, like you can describe it, but it'll change somehow. So that's yeah. I thought that was interesting.

Shana:

Yeah, I'd say research too. I forgot to touch on that as well Like it's kind of closed a gap with the number of hours that we have to spend towards researching, and that could be across any subject matter Case studies, brand information, competitor information, competitor analysis, keyword research, research, media tactic research like oh my gosh, I could go down rabbit hole it's in the middle of all these things.

Dave:

Now you know, that's the thing, yeah yeah um, sorry, I want to interrupt, but yeah, it's like it really is, like it's, it's a, it's an integrated piece of you know, like our research stuff. It's. It's huge for that now so we talked a lot about how the the tools.

Dave:

This is a tool that's it's huge for that now. So we talked a lot about how the the tools. This is a tool that gets integrated. It's um compliments the mission, it's not the mission. Right, we're able to program a brand framework inside of a GPT, a custom GPT, and that can have a brand's voice that we generate, that we make Um, we let's say, we write you know, it's a tomato sales business and so, like the tomato sales business has a voice and they have their customer and everything, and we have all that framework. We train the GPT to know those things. It sits there and waits, just like any other tool in the toolbox for us, and we go to it and we say, okay, this needs to be produced, x needs to be produced for this, for the tomato business. Inside of their voice, it will do that for you.

Dave:

And so I think a lot of folks not in a nefarious way, but more wanting to get things done they're like, well, why do I need? If I can do that, then why do I need you guys? Or why do I need somebody writing or creating or you know whatever you know? If you do that, then what it does is it changes. It doesn't eliminate our role as an agency to help a client. It changes a lot of us into more of a directorial position where we have a you know, we have a team that works on whatever the task is for a project.

Dave:

We have that GPT or that brand framework that we can keep there and then we go in there and we're more overseeing what's going on with the information that comes back at us. Does it actually work? Is it too long? Is it too many words? Is it this or that? Like little tiny tweaks happen there and then we steer that and we decide what else we want to prompt, put together and it just creates it actually creates the pieces of the puzzle for us. So it saves that jump, but it creates the pieces that we put together and we still there has to be somebody still putting this together. So I think that's a lot of with like a campaign, whether it's on the creative side or account side. That's. It's huge. I mean it allows us to actually be more efficient for our clients and do more work.

Shana:

Yeah, and I think, from both a marketing director standpoint and that internal organization and then also agencies, it's going to also help eliminate burnout a little bit over time, because I think the things that have been going on and evolving over the last few years is there's so much content needed A lot of times you need an army to create it.

Shana:

Much content needed A lot of times you need an army to create it. So you know what this does is it's going to give us the capability to have a reasonable, you know, staff load on both the marketing director side you know that internal marketing staff on the agency side. And then we have to be even more efficient or even better at what we do. So every single person that you bring on to the team needs to be an expert. They need to be, you know, excellent at their craft and, um, you know the. The tools are going to help us be able to be more efficient, eliminate that burnout, I think, over time too, in the industry, and then just be more efficient with the number of people that are on on the team.

Dave:

Right, right, yeah, no, I agree. Uh, it's, it's again, it's all this just goes back to, goes back to this changes the landscape. It doesn't eradicate it Like it's. It's a, it's a tweak to how we do business. Um, that's what all of this is. The internet was uh text messaging was.

Dave:

You know, smartphones were like. All of this is. These are pieces in the progression of humanity and communication that happen. This is sort of how it evolves.

Dave:

One thing I was going to say anybody who uses it the way we do and as a tool in toolbox, like we said, there has to be a little bit of a framework inside of an organization. Framework inside of an organization. That's one of the things I wanted to touch on. Quick is we all discussed as a team having rules around how we use it. We did touch on this a little earlier in the episode but, like we talked about, we communicate with the client. We let them know whether or not we're going to use X, y or Z. You know we may say we have these tools and we use, we utilize this.

Dave:

You know an AI style tool for something, um, but we were transparent about it. Um, we drafted a, a constitution of sorts for 1123, an AI constitution that gives just basic rules for the staff on what to do. You know. You know you don't use anything that we haven't actually approved at the operations level. You know the senior, the C-suite folks or whatever. This is an example for anybody who's a business owner. Make sure you know who's using what and make sure you're okay with it, and if you're not, there needs to be rules around that, so your employees just don't go use whatever tool they want to use, Integrating it into your onboarding even when you bring in a team member.

Shana:

I mean now. This language will be something our COO integrates every time we bring a new team member on. You know they need to know how it's when it's okay to use AI and when it's not.

Dave:

Yeah, we have some stuff that we I mean we generate. We created some. I was going to say generate, that's a horrible word to use.

Shana:

It's like the machines we actually did create you know forms that the team can submit.

Dave:

Everybody gets these forms. They can submit them to the staff, to the operations staff, and say like hey, I found out about this tool, I found this does this. Here is the benefit, here's what it costs, here's what I think we can do with it, or you know whatever all the parameters around it, and then that goes up for us to approve. I would honestly highly recommend to anybody who owns a business or is in a position to make these decisions do that.

Shana:

Yeah.

Dave:

Because that's going to help. You know you don't want to have people be like did your employee make this on whatever, or did they do this? Or is this copyrighted? And then you're sitting there flat footed going. I don't know.

Shana:

Yeah, you need to have like an agency login or, you know, brand login.

Dave:

And then those are the ones, just to prove it.

Shana:

Everybody has seats within that, you know, just actual login, so that way you can have all the historical around it too, right, yeah, I think that, and you know that's something that I think could easily get away from people. You know you're just going on about your day, people are busy. I mean we had some people on our team that were signing up for their own independent, you know, subscriptions whether it was a free version or not, and testing it out, and you know we didn't really think to go. Hey, did you all of a sudden you know you worked here for two years but did you all of a sudden this month decide to use chat GPT? So you know you're not going to think as a business owner to question the staff. So having your rules set, I think, is important, hey business owner to question the staff, so having your rules set.

Shana:

I think is important. Hey, everybody guilty over here. I did a little bit of that.

Dave:

I did a little trying, a little dabbling, before we got everything pulled in so guilty as charged. I think we all kind of did some the curiosity is just too good sometimes to be like I wonder if you know this thing does this thing. Don't lose that curiosity, but just be transparent with people, the folks you work for and with, and your clients. I mean, I think that's about the discussion here. I mean I think we covered some great things. When it comes to AI, it's always going to change, you know.

Dave:

I would say, looking at it with you know, cautious optimism is probably the best attitude.

Shana:

I think that's a good position. Yeah, good approach. Yeah, definitely yeah, cautious optimism.

Dave:

Yeah, okay, anything else? Parting shots from you. Any comments?

Shana:

No, I would say from somebody who, as you would call, is not an early adopter. Also, not being scared to test and trial within all the different options, because I did realize, you know, it's a lot easier to use than maybe you know. I was thinking ahead of time and both you know, gosh, I just think on the client side there's so much value in the work we're outputting when we're able to utilize it as well. So, yeah, not being scared to, I guess, use it in general. You know, don't be like me, I don't know.

Dave:

And firing an agency because AI exists is about the same as firing them because a Photoshop exists.

Shana:

Yep, there you go.

Dave:

Makes about as much sense. You still need somebody using the tools. Okay, so that is that, everybody. Thank you so much for listening to 1123 Presents. You can find more information about us on weare1123.com. We also have socials all over the place, of course, because we're an ad agency and we're out there. So Facebook and Instagram and TikTok and all that stuff at weare1123. And that is that. We will see you next time. The views expressed in this podcast are intended for multiple purposes some informational, some educational and some entertaining. They should be regarded as thoughts, not hard advice. The opinions of our guests are their own and do not necessarily reflect the views of 1123. 1123 presents. Copyright 2024.