...or how we taught robots to build furniture and roast our marketing ideas
Let's cut to the chase: AI is EVERYWHERE. Your grandmother's using it to write killer Instagram captions, and your competitor is probably using it to convince their client that virtual reality bingo is the next big thing. But while everyone else is busy asking ChatGPT to write their emails, MAS has been taking AI down paths that would make even the most tech-savvy marketers do a double-take.
Depending on who you ask, it's either a brilliant time-saver or a digital rabbit hole that will s u c k y o u i n. For Senior Creative Strategist Maya Smith and Art Director Shane Carey, two creatives working at the crossroads of strategy and experience design, it's both—and neither. Curious? Dive into what they had to say:
For us, we didn't just adopt AI—we've been soft-launching a situationship with it for a while. And now we're ✨posting✨.
Spoiler: it's chaotic, hilarious, and oddly human. (Like, "should we be concerned?" human.)
Beyond basic prompts: the stuff nobody else is talking about..........................yet ;)
While your inbox is probably full of agencies claiming they're "AI-powered" (which these days means they've figured out how to use ChatGPT to write subject lines), we're building custom AI systems that make our creative process feel like we're working with a team of digital savants with questionable sleep schedules.
Shane recently went full mad scientist and built a custom LLM integrated with 3D modeling software that automatically generates 3D furniture models. Yes, you read that right—he taught AI to design furniture. Not because we needed more places to sit, but because it's revolutionizing how we approach physical installations for brand experiences.
"I started treating the AI like a creative intern that needed guidance rather than a vending machine for answers," Shane explains.
"At one point, it started calling me 'Director Shane' and behaving like it was on my creative team. That's when I knew we were onto something special... or that I needed to go outside and touch grass."
Meanwhile, Maya has been creating AI focus groups that would make traditional market researchers weep into their survey data:
"We built an AI persona named Algo Alex—a perpetually unimpressed developer who's seen it all—to evaluate our tech activation concepts. Alex absolutely D E S T R O Y E D our first round of ideas, providing brutally honest feedback no real focus group participant would dare to give. It was like having that one friend who doesn't hold back when you ask if an outfit looks good."
Of course, we're not replacing real human research—far from it. Our methodology remains deeply rooted in authentic human voices and experiences. What makes our approach powerful is how we've integrated these AI focus groups alongside traditional research, creating a complementary system that leverages the strengths of both. Real human insights provide the emotional depth and nuanced understanding that no AI can replicate, while our AI participants offer unfiltered perspectives that add an additional dimension to the rich human data we collect through our research processes.
But how did we get here? How did we even land on this prompt? Did we wake up one day and choose digital chaos?
The art of prompt design: it's not what you ask, it's how you ask it
If you've ever shouted "THAT'S NOT WHAT I MEANT!" at your screen, you know that good AI output starts with better prompting. We're learning to treat prompt writing like a design discipline in itself—clear, iterative, and with an eye toward outcome.
“Most people treat AI like a psychic barista,” Shane says. “They walk up, mumble something vague like ‘make it cool,’ and expect it to hand over a perfect vanilla matcha brand strategy with extra oat milk. But AI isn’t reading your mind—it’s reading your input. Once I started treating prompts like recipes, giving it context, structure, and intent, everything shifted. Now I can guide the system to deliver exactly what we need, not just whatever it thinks we meant.”
Maya's team has built a prompt library that's become our secret weapon. "We found that adding specific constraints actually leads to more creative outputs. Telling the AI to 'poke holes and tear apart our thoughts’ gets far better results than just asking for 'never-been-done-before ideas."
A peek into our approach includes:
Layered prompting - Starting broad, then refining with each interaction
Constraint engineering - Strategically limiting options to force creative solutions
Persona shifting - Instructing the AI to adopt specific viewpoints (like "think like a skeptical Gen Z consumer")
Format manipulation - Using structural cues to shape output in exactly the format we need
AI isn’t replacing creativity—it’s widening the runway
"We're just getting warmed up," Maya insists, typically while caffeinated beyond reasonable human limits. "The AI capabilities we'll have next year make today's tools look like Clippy from Microsoft Office. But make no mistake—these tools are enhancing our human-centered approach, not replacing it. The richness of real human connection remains irreplaceable in our work."
Our innovation team—which includes several people who can explain the difference between machine learning and deep learning without putting you to sleep—is constantly developing new applications:
Space planning tools that can optimize traffic flow better than your most organized friend
Content systems that adjust messaging based on real-time audience reactions (goodbye, awkward presentations where nobody laughs at your jokes)
Predictive ROI models that can tell you which elements of your activation will deliver the most bang for your buck before you spend a dime (like having a fortune teller, but one with an MBA)
We've even started training an AI to identify when another AI is trying to pass off mediocre ideas as innovative. It's getting scary good at it. (AI-on-AI beef is the future we didn't see coming.)
The future is being prompted, not predicted
So while the tools evolve and the prompts get more s o p h i s t i c a t e d, our goal remains simple: stay curious, stay scrappy, and have some guts. We've gone from basic text prompts to custom LLMs that build furniture. From cobbling together marketing copy to creating AI focus groups that deliver brutally honest feedback that enhances—never replaces—our rich human insights.
The AI landscape is moving at breakneck speed—what was cutting-edge yesterday is table stakes today. But that's exactly what makes this moment so e x h i l a r a t i n g. Whether we're workshopping prompts at a 10 a.m. brainstorm or texting each other at 12 a.m. about some wild new feature, we're constantly exploring what's possible. (Sleep? We don't know her.)
And that's perhaps the biggest lesson in our AI journey: the magic isn't in the models or the algorithms—it's in the mindset. It's about approaching these tools with equal parts technical understanding and creative audacity. It's about seeing AI not as a replacement for human creativity but as a collaborator that pushes our thinking in unexpected directions. Sometimes into a digital brick wall, but often toward something brilliant.
As we look ahead, we both agree, we don't just see more efficient workflows or fancier outputs. We see fundamentally new ways of approaching creative problems—ways that blend human insight with computational power. We see opportunities to create experiences that would have been impossible to imagine, let alone execute, just a few years ago.
The future of creativity isn't being searched for or waited upon—it's being prompted, one conversation at a time.
xoxo shaneforsure + mayapapaya