AI is everywhere now. And while late adopters (like those in Part 1 of our series) are just starting to explore the possibilities of this new technology, early adopters are light-years ahead and are using AI in sophisticated ways to improve strategy and efficiency.
Here at Clariant Creative, we have several early AI adopters on staff, who have extensive experience using the tool in smart and successful ways that produce better outcomes. They’ve got deep insights into how you can take your marketing team’s AI journey to the next level.
Keep reading … or watch the video below!
Will Redick is a Senior Inbound Marketing Strategist with Clariant Creative and Rachelle Koenig is a Senior Content Strategist. As the driving force behind their respective clients’ strategies, Rachelle and Will have a lot in common … including a passion for AI and technology. (Rachelle does not share Will’s passion for pickleball, but she is deeply supportive.)
Both of them grabbed onto AI with gusto when it first appeared on the scene and have been using it steadily ever since. So, we wanted to know what got them so excited, how they’ve applied AI, and what lessons they can share with other B2B marketers.
We got the conversation on video, moderated by our Inbound Marketing Associate, Liza Park. Check it out, or read the transcript below, which we’ve edited a tiny bit for clarity.
For me, it came from necessity. So, this was like back in 2019, 2020, I was head of marketing, truly like the solo marketing guy for a boutique sales training firm. And that came with wearing a ton of hats. So, I was juggling multiple marketing agency partners. I was leading sales enablement, content creation, managing the BDRs. I was acting as the whole marketing department and just like doing a ton of things. So, I essentially had to operate as a five-person team in like one person.
So, when I started seeing AI tools — like at the time it was called Jarvis, but I think now it's called Jasper — pop-up that could help with like expediting content creation, I jumped right in. And it wasn't necessarily because I was like chasing hype or anything like that. It was because I needed leverage to be able to operate more efficiently.
And so later when ChatGPT came around, I realized that it wasn't just about automation and making things faster. It was a true ideation partner. It reminded me of being back before I was the head of marketing or head of marketing for a sales training firm, I was I worked previously at a marketing agency. And so it brought me back to kind of sitting all in one room with like a bunch of other marketers and like, you know, having a brainstorming session where you're like slapping ideas on the wall and seeing what sticks. But I could do that by myself and with AI and it was just like, holy smokes, this is so cool. And you know, without burning time or budget. So that's really for me kind of how that evolved that journey evolved over time.
Well, I've always been one of those early adopters of technology. So, when I was a kid, and I'm going to age myself, my dad actually was the programmer for those giant IBMs with punch cards. And so I got to go into his office and watch him while he worked. And so our family always had technology with some type of computer. I built a few before I went to college. And so, you know, as over time, I've always been really interested in new software, new technology, etc.
And so when Jarvis came out, I did just like Will, I did play around with a little bit. It was okay. But when Chat GPT came out, I got on the waitlist, like within the first week and had to wait on the waitlist and then started researching it as much as I possibly could. And then when I got an account, I just started to play around with it in my free time. Just I wanted to see what it could do and then started to make a case several years ago to bring it into the agency here to help us become more efficient.
There are so many things in mind.
Okay, so just to first off, like I use it in my personal life all the time and I've used it in my personal life from the very beginning, like almost like in tandem with, you know, how people are just constantly Googling things. I use it in tandem with that provide me step-by-step instructions. My wife is pregnant right now. So it was like, you know, tell me what's going on with these symptoms and all this. It's just like crazy how the use cases for AI, Chat GPT have just evolved.
But specifically with work, I just, I have a few. The first one is with reporting, you know, as a marketer and as a strategist. You know, I do a lot of work with HubSpot dashboards, quarterly reports and things like that. But when you have that data, data by itself is just data, right? So, I actually, I use ChatGPT to help frame the story behind the data. I'll share the performance metrics like past reporting language, the campaign context and things like that. What were our goals and give it all the context that it needs and then have the kind of comparative data and help kind of frame takeaways or spot opportunities, things that we might want to tweak or change. So that's one area in the reporting realm.
And then, you know, other ways that I also use it is kind of like strategic framework building. So I'll use Chat GPT as like a ... sounding board to help organize something like a multi-channel campaign, webinar series, or a sale sequence. I'll input things like, you know, what we've done in the past, what those results were, what we're planning to do. The key for me is giving it all the context that it needs when building out any sort of strategy. And then I ask it questions as I'm doing it. So like, what are ... where are the gaps in this? You know, and kind of let it surface this skeleton strategy that I then take in refine.
So, and then another one is kind of more content related another use case is preparing for something like a SME interview, a subject matter expert interview, where you're having a conversation with somebody who's an expert on a given topic to, you're gonna take those insights and turn it into a piece of content like a blog or a content offer or something like that. So when I'm preparing for something like that and I'm not an expert in that industry or that niche or whatever it is, I help, I use ChatGPT to kind of give me, orient myself so that I can ask the smart questions, you know, to kind of brainstorm a thoughtful strategic approach to asking those questions. So I provide things like, and for me it's always about priming those prompts. I give things like persona details, industry themes, what the goal of the SME interview actually is and the resulting piece that is going to come from it. What are our goals for that piece? And that helps me to really refine the approach to the interview with SME. So those are three different ways that I use it in marketing.
So, I'm working on a big project to develop use cases for the agency. And right now, the list of potential use cases that we've all come up with is just under 40. So that ranges from content creation to SEO to website optimization and also a bunch of internal processes. The way I like to use it, it really depends on what that use case is.
So for example, I like Claude for more robust and usable content. Chat GPT is really good for getting into the details. Partly it's because each tool has a specific capacity for length of content it can actually work in, so a word length.
So, before I even start looking at doing anything with AI, I have to think about what tool am I going to use and am I going to get what I want from it?
And then I work in and out of AI. For example, I may want to optimize a generic piece of content just to be more relevant. Maybe I want to have it be more middle of the funnel for buyers who are looking for more prescriptive information.
So, I'll start priming and ask it some of the questions like what Will was talking about, where it's like, what are the relevant best practices for this topic? What are some pitfalls for this topic? What else do I need to think about? Look at the output, then look to see is there anything that's not needed or relevant and give instructions to the AI tool not to include that.
And then I'll feed it whatever the existing content is that I'm trying to update and ask it to compare with what it actually, the results it produced. And then give me a list of recommendations. And the reason I do this instead of just having it do the work is because I want to make sure that I can manually trace where any big change comes from. So that way it's not lost in some kind of an updated version. And also because sometimes AI does hallucinate, it gets a little … over ambitious in some of the information it provides. So, I want to make sure that I'm actually reviewing and saying, maybe not so much. And then once I review that and ask for rewrites, then I'll decide if I want the tool to make the updates or if I want myself or someone else to make them. And then again, there's lots of rewrites through that process followed by a thorough review. So, in and out of the model is really helpful as well. Still saves a ton of time.
When I think about the pitfalls in terms of using AI, feel like, you know, I've seen instances where it's very clear that somebody used AI and they just like copy and paste it and just took what AI gave them and that's that.
So, I think the big pitfall here is thinking that AI can do your job for you. So, you know, people fall into this like, trap of submitting a vague prompt and then expecting like it's AI is going to know and have all of the context to be able to give you exactly what you need. But that's not how it works. If you don't provide clear context, like all the things that Rachelle and you and myself have been talking about throughout this conversation, you know, depending on what you're trying to generate and what output you're trying to generate; the context that would be needed in order to generate that output.
So things as an example, things like persona data, the business goals, the campaign goals, if you're generating something like an outline for a blog post or whatever, tone guidelines or things like that. So, if you don't have those things, you'll get very generic, unusable output.
And I think another pitfall is over-reliance. So for me, I would never send AI-generated work directly to a client. AI can, the way that I think about is that AI can get you like 60 % of the way there if you're priming the right way and you feed it the right data in the right context to in order to generate a solid output for you and it can get you there fast. But the remaining 40 % is really dependent upon the user it's dependent on your strategy your knowledge your — the nuance and your quality control. So in my opinion that's where your human expertise has to take over and you can't be too overly reliant and again just expect something beautiful to pop out and not require any human touch.
So, the idea here is making sure that people are willing to adapt to the fact that AI is here and it is usable.
It can pose a problem for folks who feel like they're not gonna have any opportunity to do what they know how to do, right now, like for example, writers, designers, et cetera.
So the idea here is you really have to learn how to tell AI what to do. There may be less of a need for someone to write from nothing and create something, but there will absolutely always be the need for human intervention. Somebody is always going to have to prompt at some point, unless someone tells me otherwise, unless AI really kind of produces or becomes more advanced. And then we're in a whole other category of things.
But really what we need is folks who know good writing and or know good design and can discern it and then can tell an AI model how to get there and what it's looking for. That's really important. And that leads to another pitfall, which is replacing AI with experience or replacing, I'm sorry, experience with AI.
So, you know, Will and I have been doing what we do for a long time. So we know what good looks like.
And the thing is, it makes me — as much of an early adopter as I am, it makes me nervous when people immediately go to AI to learn how to do something. It's one thing to learn how to do it or to try and practice it and et cetera. But when you're producing content or you're producing something and you've never done it before, you actually don't necessarily know if you're producing something good.
And so it's really important to know how to do it and then utilize AI to improve doing it. Because otherwise you could actually, we can get into this place where we're actually producing garbage and not even recognizing it. So that's why that's the push. Like within the educational realm and you know all in schools and colleges and etc. You have to know how to write first in order to know how AI can improve your writing.
So that's really important for people to think about. And then of course, what Will said about priming and prompting — garbage in garbage out. That has been the rule about technology since tech, you since people started programming, you give it nonsense, it will give you nonsense.
I would say start with the strategy and not the prompt. So, for me personally, yeah, I will sit there and compile all the things that I know I need to solve through the prompt first, like through the output that I'm trying to generate through Chat GPT.
So, you use it for so many different things, but it's like, what does success look like? What does success look like from the output that you're trying to generate from this prompt?
And then you kind of build the AI prompt from that. You prime it with inputs, obviously varying depending on what you're trying to accomplish, but in the marketing realm, things like campaign history, samples of tone. I've played around with some of the other things like, you know, try this in like a Ryan Reynolds-Deadpool type of tone and things like that just for fun. But, you know, having that idea of whatever your client's tone is or what your organization's tone is, and brand voice.
Things like priming it with persona summaries or even examples of past work, all of those things, building your prompt from that.
I think the way that I think about it is, the more that you teach it, the smarter your output becomes. You're even able to, within your own AI accounts, you're able to like call back upon like, hey, remember that output that we generated previously for this campaign? Like you can, you can do that and reference that when you're thinking about the strategy that we're trying to, you know, to accomplish here. So the big thing is, yeah, the more you teach it, just like we were talking about before, the smarter your output becomes.
And the last thing that, you know, I think about when it comes to, to this is, is always finish the job yourself. You can use AI to generate that momentum, but the human being behind the computer or the phone or whatever it is needs to have full ownership of what the final product or output is. So it's still your name on the line. It's not the AI tools.
I would say kind of reiterating what I said earlier, just be careful not to replace your own knowledge with AI. You need to know good work when you see it and that requires some level of work. And to Will's point is making sure that you're as involved as possible.
I think the other thing is that people tend to take prompts and copy and paste them and assume they're going to work every single time. And the thing is to remember is AI is progressing so quickly. Like how AI worked and how easily it could discern intent, it couldn't in the very beginning and now it can. And it's getting increasingly smarter as more and more people add prompts and primes to just the general AI model.
And so as a result, this kind of solidifies what I had said earlier is you really need to know to Will's point what you need, what you're trying to accomplish, what your strategy is. But you also need to be able to think about what prompt or what priming do I need to do right now for this project? Always reuse your prompts if you've done the work to figure out a whole bunch of information. AI models work better when it talks to itself more and more and more.
Last thing is the one of my biggest ongoing concerns with AI, believe it or not, even though I'm an early adopter, is the bad actors who might use it nefariously. So, we can think about that on a macro level, right? People who are doing things negatively, you know, globally, let's say, but it can also happen on a micro level, which could be worse.
So, be the person that's using it responsibly, with integrity and appropriately because that's just creating more positive energy, versus creating this practice of not utilizing it in a way that's authentic. And that's creating a blend of content that's a blend of human intervention and technology that's really making it easier and more efficient to create.