SPECIAL SERIES: How to Avoid AI Slop in Brand Creative with Michael Horton, Co-Founder of Slate & Adam Creeger, CTO at Slate

24:14Michael Horton:I don't know. And Meta, you like you saw like how like content quality like. Affected like billions of impressions in a feed. So like what did you learn there that about what makes like AI assistant content feel more humored, human versus generated? We're brought to you by Slate today. And honestly, if you run social for a brand, listen in for like 20 seconds. How many tools are you using just to make content? Because most teams are at five or six apps. And you're still chasing logos in some Google Drive folder nobody's organized since 2022. Slate is one content creation workspace. The NFL, Amazon, New York Times are on it.

25:05Michael Horton:The Pistons Lily dish cap cut for Slate. video and image editing ai clipping approvals brand assets all of it slate teams.com links in the show notes back to the episode yeah like you know when i when i was leaving meta was around

25:23Adam Creeger:the same time that like ai created content was really getting into the feed i think the thing that was consistent throughout the content I observed, both from live and reels. It was just the importance of authenticity, the ability for content to connect with people, even if it's from a brand. That's what made live so engaging, was that it was, at the time, just this really great way to see the sort of the uh the real person behind the you know the it's the instagram versus reality like the irony is like it was all about instagram was about this curated version whereas like live gave this authentic um view into um someone's world and i think over time you know that permeated into like other forms of video especially like short form video where like the the best content is authentic and that is you know truly like more true than ever with with ai um and it doesn't just mean you can't use ai to be authentic like there are um people out there using ai entirely to tell authentic stories right but it's it's just that i think there's authenticity and it's a bit of a cliche but there's always like certainly for the last few years always been this common thread between high quality content.

26:53Adam Creeger:And the other piece, going back to what we talked about earlier, is just the level of care and passion put into the content and the level of detail. We talk about this a lot at Slate. The way we see AI being used is to get us, whether it's coding or whether it's content creation, to get our users to 80 % quickly get through the mundane that first 80 piece quickly and really turn that last 20 percent into actually more than 50 percent and spend a lot more time like getting the details right really creating something that's like curated and polished and like as close to to perfect as possible um and that's what you know i've seen have the biggest impact is that like how how much time can you tell someone who has spent that time like getting things to like grazing the quality bar because they got to that 80 faster so they invested some of that time into that last mile and that is you know that's what i started to see become the differentiator i i want to know

28:10Michael Horton:what what you both and your thoughts of what is that that 20 like what is that 20 that like humans are doing versus the 80 % that AI will take off your plate? What does the 20 % look like?

28:26Adam Creeger:Yeah. I mean, I think from a content creation standpoint and it's sort of my mind immediately goes to video production. That's where my background is personally. And it's also like one of the main uses of Slate. But the same applies, as similar things apply to written content, image content as well. But it's kind of easier to map out from a video perspective. There's a lot of tedious work, a lot of times tedious and just time-consuming work that leads up to creating a piece of content, even a short form piece of content. And some of that could include things like finding the right media, finding the right spots in longer form media, you know, like, like, I don't know, this is probably something you do with these podcasts, but I'm sure you're familiar with, or someone on your team's familiar with scrubbing through, you know, content to find the best soundbites.

29:24Adam Creeger:Oh, I'm so familiar.

29:25Michael Horton:I'm so familiar with that. And before AI, it was the worst thing in the world. Right.

29:34Adam Creeger:And AI can help you get there, but it's not, you know, our approach to this is AI can pull out moments and can help you or help you find the moment that you're looking for much faster. but it's not going to be the, or it shouldn't be if you care about your voice and your craft and, you know, uh, the overall quality of the content you're creating shouldn't be the one making all the creative decisions on exactly how to set up that soundbite, soundbite, like where to cut, where to not cut, um, you know, how to stylize it, how to make it match your brand. Um, so, you know, it can help you find the right media.

30:12Adam Creeger:It can organize all your media. It can kind of to help you find the right points in that media. It can even make some basic cuts for you to set things up. And similar kind of processes could be applied on written content. But if you're not spending, again, that last sort of 20 % applying your own creativity and your own style and your own voice, making sure that that's matching exactly what you want that outcome to look like, then you are going to be creating very generic content that probably is not, it's going to get wrapped up in the AI slop bucket that we keep talking about. So yeah, I mean, I think there's a lot of workflow pieces that AI can help with that give you more time to focus on that last 20%.

31:05Michael Horton:Yeah. I mean, I, I'm from my experience too, though, it's like, AI could find clips, but at the same time, someone has to know like the audit, what the audience would actually react to versus not like, and that's the problem like I have with podcasts too, with for myself is if I, if I don't go in there or someone that's a marketer doesn't go in there and looks for the clips that would react to marketers, it's just going to be like clips that are going to get very little views like the views that will be it's like those hot takes um like stop starting where any like starting at a cliffhanger um things that like ai doesn't think about that it just would be like oh these are could be good clips and like if you're not a marketer or you're not a sports person or whoever your audience is it's hard to like pull out and say like my audience would actually like that and that's like the biggest issue i have with like that's why i don't like ai i never say ai is perfect because like it gets me to the place where i could find my clips but at the same time like i have to go like look and find the clips otherwise if someone's not a marketer they're gonna give me five clips that i don't i wouldn't react to um so yeah and then i also think that 20 i've got

32:28Adam Creeger:I agree with everything Michael said. I also think there's this like very practical things that you should be doing with that 20%, which is like, check your captions. Like don't just trust the AI generated captions. The number of videos I see online where the author has just clicked the caption button and just trusted whatever has come out and posted like simple things like that, fix them, even tweak the layouts of the captions so they flow well and make sure your cuts are great, make sure use that time to experiment, try five different variations of an edit instead of just one, right? Really, I think it comes down to just craft like go deeper into what you can be doing with that time.

33:18Michael Horton:yeah and like safe zones and stuff like that where people don't think a lot about

33:26Adam Creeger:yeah that's a great great great uh feature drop for a slate there so one of our

33:31Michael Horton:most beloved beloved features my goodness it's the worst thing if you like because i used to like because i create a lot of memes and like the worst thing or creating the meme is like you put it on a platform and then you have to like go like size it to like make sure it like fits in to like the screen and if it doesn't size properly i have to like redo the whole image again and it's just like a painful process to not have safe zones applied in the product and that's why i was saying like other tools like like cap cut um doesn't do that like um canva doesn't do that it helps you like get maybe like half the way there but you still have to like know what it looks like on platform or not.

34:16Michael Horton:Yeah.

34:17Adam Creeger:If you're listening to this and that's a problem for you, then Slate is a great way of solving that problem. I also just wanted to like go back to what you were saying about, you know, like speaking. It takes you to find the clip that's like speaking to the audience and lay it out in a way that's going to connect with them or a format that you know is going to, you know, catch their attention. Like that's totally true, obviously. And I think that's part of our belief on this is it's always going to be better for the social media manager, the marketer, whoever it is that understands their audience and the goals of their content, what they're trying to create, to be in that loop somewhere in that process of making those decisions.

35:03Adam Creeger:But I also think right now, some of that goes back to the stage we're in with a lot of the tools that people are using are built to optimize for, for like pretty generic things. Like they'll say like, Hey, here's the most viral clip, you know, but they're not thinking about, well, I'm not looking for the most viral clip. I'm trying to like, you know, one, what do you actually know about what being viral is going to, you know, what actually is going to be viral, but I'm actually looking for something that speaks to my audience. You know, I'm looking for something that follows like, okay, on this platform, This is the type of authoritative voice I'm trying to like execute on or, you know, the strategy that you have as a brand, as a, as a social team, as a professional is different than like just someone trying to create, you know, the most viral clip they can.

35:53Adam Creeger:And most of the tools just aren't set up in a way that you can do that. So that's another thing that I think makes, we'll see tools continue to evolve. And I would say Slate's leading the way on that is making sure that when AI is trying to help you get as close to 80 % as possible, what that 80 % looks like is actually taking into account what you as a brand care about. It's not just like in a vacuum, here's the best clip. it's what's best for your goals as a brand, like for your voice, for your output, for this particular scenario, what's your past behavior, what's your, you know, kind of past preferences, taking all that stuff into account.

36:33Adam Creeger:It can help you get closer to your actual desired outcome, not just a generic, like good outcome, you know, and that's where I think a lot of the tools are sort of missing the mark as well. Yeah.

36:44Michael Horton:And I think there's like two ends of the spectrum too. and I want to go into the middle ground of it. There's one end of the spectrum where the tool is built for a creator and a creator has more leeway with brands. They don't have to care about IP and all the stuff that brands need to care about. That's one end of the spectrum. The other end of the spectrum is everybody is using this template, which means that it's average, which means it's boring content. Um, and I think like, I want to go into like the AI guardrails because like the AI guardrails could help you like produce really boring content or it can help you on the other, just be like in addition to creating great content.

37:34Michael Horton:And I want to know where that thought process comes to both of you when it comes to AI, because I think without the correct guardrails, it could end up being maybe not sloppy, but generic, boring content that's overly brand safe.

37:57Adam Creeger:Yeah, I mean, I can start on the way I think about that, at least.

38:05Adam Creeger:I feel like we're spending this whole podcast just hyping up Slate, but obviously we think we're taking the right approach to these sorts of challenges, and it's something we talk a lot about. But, you know, I, you know, I think the guardrails that are in place should also come from your brand. You know, it's not a generic guardrails. Like this is the sort of AI guardrails that are applied. Like you shouldn't do this. You should do this. And that's being those guardrails or those rules are being generated based off of the LLM and the aggregation of every piece of content in the world. It's being the guardrails that are being put in place are being created based off of your own brand.

38:55Adam Creeger:And going back to things like those, you have a 50 page PDF on what your brand should look like. and maybe you have a 50-page PDF on what your brand should look like on social media specifically and voice and style and everything, that can give you pretty good guardrails of where to start. But other brands may have much looser restrictions. And they might say, we actually don't want AI to be telling us to not use this color or not use this font or use a different voice because we're trying to be super raw and loose on social. So, you know, I think it goes back to setting up whatever AI tools you're using in a way that follows your own guardrails and how strong or how loose those guardrails are should depend on your goals as a brand, not just what AI thinks guardrails should be.

39:52Adam Creeger:So that's kind of the approach we're taking with it. And obviously we think that's the right approach for the future. I also think different types of content require different types of guardrails like i think like written is easier to apply brand voice

40:11Michael Horton:guardrails video is a little different because there's could be a creator involved or a person involved and you don't want to overly engineer brand voice guardrails to someone who is a human creator that has their own voice so i think like that's why i think it's channel content format specific what guardrails should like end up being in your ai software that you're using uh yeah i

40:40Adam Creeger:think i i think you know guardrails are you know can be a form of constraints and the right kind of constraints can really make you more creative not less creative um i i kind of the analogy i think of is a bit like gps right you can see gps in two ways you can either see it as the thing that tells you exactly where to go. Or you can see it as the thing that allows you to actually go wherever you want. But when you turn it back on and when you use it, it will always get you back to a good place. Right? And, you know, I see like, and not every depends on the guardrails and the brand guidelines themselves, but, you know, brands can use this with AI and with like their creators to encourage creative freedom, but still yet have like still have confidence that there will be some consistency in the output.

41:35Adam Creeger:Wherever the creator goes with their ideas, they will still come back to a place that is kind of known and agreed upon. But it's that journey that's kind of the interesting and the exciting part.

41:47Michael Horton:I love that analogy because Rory Sullivan, who's a famous psychologist, says that Like most like GPS is a very logical, like they, like they help you find like the fastest route to get somewhere in the quickest way. But a lot of humans, like maybe you want the illogical thing. Like I want to take the scenic route home, or I want to take the route home that gets me past this sort of restaurant or like, they want like illogical things sometimes with their route. And And it's like maybe the drive, oh, I don't want to sit in traffic all the time. I want to be moving the whole time. There's like so many different ways that a GPS, but the GPS is going to just tell you, you can get here within 20 minutes.

42:34Michael Horton:But like you said, the GPS is helpful if you're like driving that scenic route and you take a wrong turn and you don't know how to get back into the main road. And so I love that, like, you need a GPS as like, you need some logic applied to what you're doing, but you also need that other, what humans want, the logical stuff. So you need both logic and illogical things to make great content work at the end of the day.

43:04Adam Creeger:Yeah. And the path to creating great content is much more about that journey than destination. And I think, you know, sort of having like a great system for capturing like these brand guardrails and guidelines just helps you take your own journey and yet still have some idea of like where the destination might be.

43:31Michael Horton:I want both your perspectives on this too, because you both have said that we're in like the infant stage of AI. um so where where do you see like video editing or video production um in three to five years like applied with ai like it could be your best guess but where do you think like how is it going to like evolve for like this space um in like the next three to five years um adam i can start with

43:58Adam Creeger:you yeah um i think it's you know obviously it's a big question and like i said a lot of the the progress and evolution here is going to come from our ability to use these tools better. You know, I see that people will still have a very fundamental role in the process. The value of creatives is not going away. I think AI will play more of a role over time. I think that we might start to see things like creating more variants of things, more personalization come into play. But I think fundamentally expectations will become higher. Expectations from both consumers and also brands of their creatives will become higher because they have more tooling at their disposal.

44:57Adam Creeger:And I think that we're going to go through a pretty disruptive period, period. But I do think we're going to get to a point where we're seeing a lot of great and innovative content being produced. Whether it's generated by AI in some cases, whether it's edited by AI, but fundamentally created by humans, I think we'll see just this mix of content in the ecosystem. But at the core of it, it's always going to be a human and their creativity because by definition the ai technology we have today is very like um it very much depends on what's come before it and to truly innovate is a human quality and i think that's become more and more apparent the one example i think about

45:49Michael Horton:right now is like how that like if you don't know like in social like that puerto rico song went viral but it was like a guy who read wrote the lyrics went to puerto rico like came up with all like the thing but he used ai to like generate this like the song because he's not a singer but he went it went viral because it was like human inputs but like ai helped get him to like the like the cat like and he described exactly what he wanted as the music like i want this and this and this and this is what the song i want to do um but it was like a part like a perfect example of like human creativity layered with ai instead of just like write me a song about puerto rico it was like yeah true experience about puerto rico and then here's a song um

46:33Adam Creeger:and and you know if everyone goes and writes a song about where they go on vacation or where they they visit i'm gonna get pretty bored of that pretty quickly so he was like he had the creativity yeah so you know it's it's not that like again it's that he had the idea to do something new and unique. That's a really good point. That is forever going to be the important piece of the puzzle. And AI is just going to enable us to do more and more creative and interesting things. I do think too, it is like a pretty good example of, maybe not a perfect example. I think a better example, I think of where things are going is going back to the parallel you were drawing earlier, Adam, of what we've seen happening with coding software, where engineers 18 months ago, many of them probably wouldn't have thought that they would be writing less code by hand or writing almost no code by hand today.

47:37Adam Creeger:But that is kind of what we're seeing happening. That is what we're seeing happening. They're writing less and less code by hand, and they are becoming a little bit more like directing agents that are writing code. But as we said earlier, that doesn't mean we need engineers any less than we did before. And the ones who are the best at writing code are still the ones that are the best at directing the agents to write code and doing it in a more powerful, more creative way. it does take a little bit of a shift in your skill set and your mindset. You have to be, frankly, better at multitasking and code switching.

48:24Adam Creeger:Maybe not code switching, but switching your mindset from task switching. But you have to be still, ultimately, a good engineer to direct these agents to write good code and understand when they don't and tell them what to do differently. And I think content creation and video editing is going to continue to go a similar way where, you know, as an editor, I won't be sitting in front of my timeline manually scrubbing through everything and finding like the part I want to cut out and highlighting it and deleting it and right clicking. and I'm doing less manual tasks. I may be directing agents or the software I'm using to do some of that, but ultimately the people who are creating the best content are still going to be the ones who have that vision for where they're trying to go and they're executing it using the tools available to them in the most efficient way they can, which is going to be using AI to sort of like we said earlier, cut out all that tedious work or just help you get to your end goal quicker.

49:27Adam Creeger:But if you're leaving it up to the AI to decide what that end goal is and what that vision for the final piece of content is, like, then you're not going to be creating good content. Even, you know, three, five years from now, you're still going to be creating it.

49:40Michael Horton:yeah i i always think of it in like a sports analogy too is like like if everybody had access to the same like nutrition plan like the same like like way to train the same gear like the the better athlete is going to come up up top and that's exactly what i think is going to event like down the road is going to happen with ai is like the better marketer the better engineer the more experienced person who's actually putting in the work to like level up their skills not only their AI skills. Um, cause right now the AI is like the nutrition could have gave like, like Novak Djokovic, like the better edge for a long time.

50:21Michael Horton:And then until everybody figured out like, okay, nutrition is like actually something that you need in sports. And then it's like, when everybody's doing nutrition, like then the best, the athlete becomes like, like the loving playing field, not just the nutrition. Um, so I think like right now people are like figuring it out like what like and the ai is giving them an edge but like the better athlete or the better marketer or the better engineer is going to come out on top eventually that's my hope i think um and um lastly i want to ask because i ask everybody on this podcast i know coming close is what's a marketing hill you would die on like um a marketing stance man a marketing

51:03Adam Creeger:hill you would die on? Hmm. You're asking two, like, not traditional marketers. Well, it's

51:13Michael Horton:good because I can't even ask you from, like, let's ask in a different way, because what's an AI hill you would die on?

51:22Adam Creeger:Well, I can talk a little bit about marketing hill I'll die on. Okay, okay. It, to me, goes back to just the love and the craft. Like, even... Like even the simplest content needs craft, love, time, like, and creativity. Like you all, you know, there is no shortcut to great content, great content. Like, and that also, you know, if it's really about marketing, it's also about the insights behind the marketing. Like great marketing doesn't just come from the end result, it comes from the whole process, right? but you know understanding your audience understanding you know what you think about what you need to say like how you need to say it all like all of that work which has to happen likely before you ever get behind a camera before you ever sit in front of a video editor before you open up photoshop any of that stuff like that is still that that is the the the really valuable piece um and i think if you're skipping that then you're really limiting how impactful the work you can do will be yeah i love that because i i i truly believe like

52:43Michael Horton:out of every way every tool every software that comes back like the fundament like like people stop thinking about the fundamentals of like the craft and they like get away from the fundamentals and you always should be working on the fundamentals and the craft to get better at those things because some things will always ring true, like human behavior is not changing,

53:07Michael Horton:like knowing what good copywriting looks like, knowing what converts. Those things will never change, but how you do it is changing, platforms, tools, AI. is like gonna but the fundamentals is like learning what a what makes a human react or not um like emotionally yeah yeah i mean i i think like the one i was thinking about is similar to

53:36Adam Creeger:it kind of you could kind of apply to both like the ai side and the marketing side but it is yeah similar to which i think is like centered around sort of like authenticity essentially but Or like, you know, that I think, I think, I'm trying to figure out the right way to word it, but basically the idea that trying to create like a viral piece of content or like a trend just to, just to be following a trend or just to be trying to create something viral, like is always going to fail in comparison to really understanding your audience and creating specifically for that person. And, you know, I think that that's something that a lot of brands even today still fall into that trap, especially brands, I would say.

54:21Adam Creeger:You know, it's obviously where like creators, individual creators, I think shine through over brands. And I think we see a lot of like the bigger brands leaning into, um, you know, like faces for, for, to represent the brand, you know, leading into, um, you know, sort of like their, whether it's their executive team or their social media team or people within their organization that, um, are creating content, you know, cause they're trying to like capture more of that authenticity where a lot of brands fall into the trap of just like following whatever the latest trend is or whatever's going viral and then trying to like replicate it, apply my brand into it, like shove it in there and then post it.

55:02Adam Creeger:And, you know, I think, um, it's not a hot take, but staying true to your, to your audience, like it's going to always win.

55:11Michael Horton:I do believe that. Cause I also think that brands fall in the trap of the checkbox. Like we have to hit this revenue goal, um, where like a creator is always thinking about like, Like, I want to create the best video for my group of people. And they don't stray really away from that if they're a great creator. It's like, I'm creating the best video for or something my audience will like. But then we fall in a trap as a brand of like, we have to hit these metrics, like social metrics are down. Or I have to hit this checkbox of we need five posts this week. Or I have the content calendar or this.

55:49Michael Horton:And we're like forgetting the end product is to create the best content for the audience. Like that's the ultimate goal of creating content is creating the best thing for the audience. Lastly, where can people find you on your journey?

56:04Adam Creeger:Well, you can find Slate at slateteams.com. You can sign up to talk to someone on our team if you're interested in checking it out. Um, you can find me on, I mostly, uh, post on LinkedIn, even though Adam was, but you can find me on LinkedIn, uh, Michael Horton. So, uh, that's where you can find anything I'm doing. Yeah. Um, likewise, like, uh, yeah, I'm, I'm actually fairly private on social ironically. Uh, but yeah, LinkedIn, Adam Krieger, uh, on threads, just Krieger, um, same,

56:40Michael Horton:same for twitter slash x as well well thank you both for coming on and thank you for like sharing like it's good to have a different perspective from like an engineer mind and a product focused mind of like where this ai ai trends are going and especially in like the marketing sense so i appreciate your time thanks so much for listening keep tuning in to hear more great insights from the coolest marketers from around the world. If you haven't ready, make sure to subscribe and follow the Marketing Millennials podcast on Apple Podcasts, Spotify, YouTube, or wherever you get your podcasts. And if you like what you hear, I would greatly appreciate you giving us a five-star rating.

57:24Michael Horton:It helps bring more marketers into our community.

57:36Thank you.