The Lacek Group’s artificial intelligence (AI) team—with members from across agency functions—takes a proactive response to keeping our work on the cutting edge of emerging AI technology. It stays abreast of evolving laws, explores available AI technologies to share that knowledge with colleagues, and strategizes the development of prototype AI tools to meet the business needs of our world-class clients. These team members gathered to reflect on AI’s challenges and opportunities.
DUWAYNE MAHLEN: AI has definitely captured the zeitgeist. In part, that can be explained as the latest FOMO response to a new technology. That phenomenon—fear of missing out—probably started with the invention of the wheel. But more recently, we’ve seen it play out in the clamor to use the cloud and then in the obsession with advances made possible by blockchain technology—nonfungible tokens and Web3.
JULIE BUSTOS: And AI is the current source of FOMO.
DUWAYNE: The problem, though, is that it can create pressure to act quickly, skipping steps or rushing them with shortcuts to push the new tech live. When that happens, it’s usually security and privacy reviews and quality assurance processes that take a hit. Worst-case scenario, that could mean misconfigured settings or bugs that lead to leaks of data or personally identifiable information.
GREG GORANSON: And the portrayals of AI in popular sci-fi movies have been fanning the flames for years. In some, AI is scary (think 2001: A Space Odyssey and The Matrix trilogy). Sometimes it’s searching for connection (for example, Her or WALL-E). I think people are just fascinated by the idea that we’ve created a machine that can think for itself.
TODD HEDBERG: Exactly. Pop culture has a huge impact on perceptions of AI. Some people are excited by it; some are overwhelmed.
SHI BU: Wherever you go these days, everyone is talking about AI. From a data practice perspective, I’m excited about the potential for AI technology to tremendously improve our productivity. But let’s remember that AI isn’t actually new; it’s just that before, its capabilities were controlled by a few people behind the scenes. The difference is that now anyone can use and benefit from it.
We’ve been using a machine-learning algorithm for some of our client work for a few years. Some would say the predictive analytics we use are an AI function. We’re not yet to the point of broad use of LLMs [large language models]—like ChatGPT and others. But some applications we use for various clients incorporate generative AI features—such as Salesforce Einstein, Adobe Firefly, AIShare.
JULIE: Ultimately, predictive and generative AI are just tools. They’re a new means to achieving our goal: building and maintaining brand devotion for our clients.
DUWAYNE: Right, it’s about how people use it. I wish more people understood that AI isn’t an encyclopedia. It’s not a thing we ask for facts—we have search engines for that. LLMs are for creating language. When I test them, my go-to question focuses on context rather than fact: Why are blueberries blue? Then I can really see if the generated language sounds natural, if it flows like human speech. One of the LLMs I tested was sarcastic in response to my blueberry question. It said, “Because they’re delicious.” I followed up with, “Why is the sky blue?” Its response: “It isn’t.” Those responses fail the context test.
JULIE: It’s important to remember that AI is really just looking for patterns and guessing in what order it should regurgitate them. So sometimes you’ll get nonsensical words and sentences. That said, generative AI has a pretty amazing capacity to be coherent. Sometimes its facts or context are wrong, but at least the text is lucid.
However, sometimes it misses the mark. Like if you ask it to generate an image of something that looks bananas. It’s likely not going to understand that idiom, so it’s probably going to build an image of fruit rather than spit out an illustration of something kooky or silly. At the moment, it just can’t nail that part of it.
GREG: That’s why you don’t ask AI to make an image that includes a person. It can get a face perfect, but it can’t get hands. They’ll have seven fingers, or the hand will be sort of mashed looking. We all understand what hands look like—we deal with them every day. But AI can only understand them in two dimensions, as in a photograph—for now anyway.
But I wish the inherent bias of AI would be more widely acknowledged. As Julie said, AI essentially scrapes swaths of existing digital information—much of which is racist, homophobic, misogynist, xenophobic, or flat wrong. We live in a biased society, and the internet is a reflection of us. At least right now, generative AI is only as accurate and objective as the data it’s fed.
JULIE: That makes me wonder—is it possible to eliminate the kind of bias you’re talking about? Who gets to decide? There are so many ethical complications.
GREG: Absolutely.
DUWAYNE: We use Alexa to set alarms—you know, to wake up or to time something in the oven. When my wife asks, she finds Alexa’s answers condescending. When I get the same response, I don’t feel that way at all.
GREG: I get in fights with Siri all the time.
DUWAYNE: Just like a conversation between humans, there’s an X factor to how each person, in any given moment, might respond. That’s not the same as bias, but it’s at least adjacent.
TODD: Some studies show that using your manners in AI prompts—like please and thank you—actually produces more detailed responses, which is fascinating, right?
GREG: Isn’t that true of human beings too?
TODD: One hundred percent. If my kid wants a Popsicle, she’d better say please and thank you! Look, I also think it’s important to remember that AI isn’t, on its own, a loyalty marketing or CRM strategy. It’s an innovation that has the potential to elevate brand strategies. Like any other tool, it’s only as useful as how it’s applied. I find myself having conversations with clients who are eager to get started with AI as soon as possible. But first we need to determine the right use cases for AI for each client—then we can create an appropriate strategy to implement in campaigns.
SHI: Yes. We really need to build people’s understanding that using AI is a skill we need to learn—it’s not a tool that operates entirely on its own. Exploring it may lead to really innovative solutions to problems we haven’t even encountered yet.
JULIE: For sure. It’s amazing how knowledge evolves. Like, at the moment, we probably don’t have the vocabulary to even describe what AI will make possible. It could be that transformational.
Maybe we’ll advance to a kind of “infinite loyalty”—to adapt the term Todd coined in his white paper on infinite personalization. Maybe in 10 years there will be enough data feeding enough computers to figure out what exactly will make you loyal, what will make DuWayne loyal, what will make each of us loyal. And we’ll be able to serve that up on an individually catered basis. That would be mind-blowing. But—and here’s the sticking point—that potential future is built on our current understanding of loyalty. My hope is that AI will transform how we operate in the world. And, if so, the future of loyalty is probably unlike anything we can even imagine today.
SHI: I imagine generative AI will help us change the proposition of loyalty. Today’s marketplace is feeling the impact of influencers on social platforms. I think that will only grow, and that gen AI will play a big role in that.
If you look at influencers on TikTok, Instagram, etc., you won’t hear them talking about loyalty programs—you’ll hear them talking about how they love certain brands. Gen Zers view influencers as more authentic than traditional brand marketing because they’re real people sharing their own experiences.
Gen AI gives us the capacity to create digital influencers for our clients on these platforms—to really advocate for the benefits of a loyalty program. Now, I’m not saying this is an idea we should necessarily pursue, just noting that it’s the kind of potential gen AI opens up, particularly to reach younger generations.
JULIE: I think the pace of innovation will only increase. That’s part of the reason why it’s really difficult to predict where AI will be in five or 10 years. We keep coming up with new things.
As Shi said, the power of influence is big because people trust human experience. So how will customers respond to digital influencers? Will they be received well because they reflect the human experience that trained the AI? How will inherent bias affect that? What other legal and ethical issues will crop up? I have so many questions.
If we get to a point of using AI to create infinite personalization in which we can identify and target each individual’s loyalty triggers, we create a new problem: What’s the differentiation? Why would someone choose to be loyal to one brand versus another?
SHI: One thing I know for sure is that we’re going to be bombed with a deluge of low-quality content. It’s already happening, and it will only get worse.
DUWAYNE: And remember, the people trying to legislate AI will almost certainly not really understand it. (We saw this play out in Bitcoin legislation.) They’ll remember I, Robot and worry about technology taking over the world rather than the practical ethical and legal concerns of a tool, not an existential threat.
GREG: Right. And one pressing concern is copyright of creative material. Already, AI has created a certain Wild West, anything-goes atmosphere about images, writing, music. Legislation hasn’t really caught up to where we are at the moment, much less looked ahead to where we might be going.
TODD: But along with unknowns and complications, AI tools will surely present powerful opportunities for new efficiencies. As an agency, that could help us reduce human hours, and therefore cost, so we could be more competitive on pitches and win more business.
At the same time, what’s preventing the client from doing some of those AI-enabled automated activities themselves? Ultimately, I suppose it boils down to that perpetual tug-of-war between quality and quantity. Even the existing AI tools have proven they can provide quantity, so it will come down to judging if the quality is good enough at scale.
JULIE: As is often the case, the winners will be those who find successful ways to adopt AI quickly. Laggards will get caught up in their established processes, unable to be agile in experimenting and sharing knowledge and new approaches across organizations.
AI isn’t a creative problem or opportunity. It’s not a data opportunity. It’s not an app development opportunity. It’s an everyone opportunity. We need to explore how we can use AI to do what we do better, which will require the varied perspectives of everyone in our agency to imagine.
GREG: Right now, AI is the shiny object in the room. Everybody wants to talk about it. It’s our role to guide our clients to not get distracted by its novelty but focus on identifying what loyalty and CRM challenges they’re encountering. And then considering if and how AI capabilities might be the right tool to address those challenges.
I don’t see AI as a replacement for human creativity. But I think it has the potential to help us be exponentially more creative because, in theory, we’ll be able to do so much more in a shorter amount of time. Setting aside my worries about copyright infringement, bias, and inadequate legislation, I believe AI will really open up opportunities we have yet to uncover. That’s exciting.
JULIE: What gets me excited is how AI is such a sea change—a cultural shift with a clear before and after—like ATMs changing how we bank, microwaves altering how we cook, and the internet transforming almost everything. There are concerns for sure. What does it mean for creative ownership? What does it mean for privacy? But I’m an optimist, so I think we’ll figure it out.
TODD: I’m excited to be a part of the AI sea change, as Julie calls it. I think it’ll allow marketers to really be optimizers for customers—delivering not only hyperpersonalized messages because we can, but because it can improve the quality of life for people no matter what product or brand we’re supporting.
SHI: I agree with you all! I encourage myself every day to not let fear prevent me from learning something different. That’s how I feel about AI. I have some anxieties about it. I have some moments of pessimism. But I won’t let that prevent me from exploring its potential. I hope everyone will take that approach.
Shi Bu, SVP, Data Intelligence; Julie Bustos, SVP, Marketing Innovation; Greg Goranson, creative director; Todd Hedberg, senior director, Digital Strategy; and DuWayne Mahlen, VP, Information Technology, are part of The Lacek Group’s cross-functional AI team. For more than 30 years, The Lacek Group has been perfecting the art and algorithms of brand devotion. We help world-class brands identify their highest-potential customers, engage them across channels throughout their life cycles, personalize each relationship for optimal long-term results, and measure the true effectiveness of those efforts.