Twisted Pair: A Case Study in AI-Assisted Product Visualization

AI didn’t play sports in high school

At Spark No. 9, clients come to us with big visions, and it’s our job to make them real. That means imagining products that don't yet exist—often for audiences that are niche, and specific.

Lately, we’ve been exploring how AI could support the product visualization process. Can it help us move faster? Sharpen concepts? Help visualize the product better?

Let’s find out.

We ran an internal workshop to generate new ideas that would allow us to start building some AI muscle. A few standouts: aroma headphones, savory spicy protein powder…

But the winner? Twisted Pair: the first retractable cleats you can wear on and off the field. Sean, our resident sports guru, was the idea guy here. But even those of us who have never laced up for the big game appreciated the idea in an instant.

At Spark, we test product concepts with more than one audience. Our clients are almost always surprised by the way something new resonates with a completely unexpected audience segment.

When thinking about audiences, it's important to ask: what problem does this product solve for them? In this case, it’s about saving money and adding convenience—one pair of shoes instead of two for practices, meets, and events. That’s not a pressing need for professional athletes, but it could be a major win for someone else.

What about competitive kids and teen athletes? Parents are often willing to spend serious money on anything that gives their kids an edge—or makes their gear more versatile. Retractable cleats mean safer movement on hard surfaces, more comfort off the field, and no more awkward stinky shoe changes in the car after a meet. It checks all the right boxes for parents: safety, convenience, and cost savings. It’s a win-win.

But just try getting your kid to wear something practical. To check boxes for kids, the product needs something more.

As the name suggests, Twisted Pair is built to be cool, innovative, and disruptive. The brand and product design needs to reflect that. We need kids and teens to want these.

We started with palette. Think acid greens, toxic pinks, infrared oranges. High-contrast, retina-burning colors that grab attention. The shoe would feature a visible retraction mechanism, leaning hard into a neo-futurist, glitchy-yet-sleek aesthetic, almost alien. (Imagine if H.R. Giger designed soccer cleats.)

(In that case, could it catch on with sneakerheads too? Maybe—at least the ones who love weird?)

But while the style matters, we can’t forget: this product also has to win on convenience, comfort, and cost.

We started AI-free, with help from our talented UX lead Jessie, just to get a general vibe for the super-cleats:

Next, we explored the feasibility of the shoe itself. If you're an experienced industrial designer at a major athletic brand, how realistic is this idea? We had envisioned a shoe with retractable cleats—such a thing exists for snow boots. Was that pragmatic here?

We do a lot, but we’re not industrial designers. So, we asked AI to help us figure it out.

The result? A 25-page breakdown of potential mechanisms. Here’s the short version:

Designing a convertible sneaker-to-cleat athletic shoe is technically complex but potentially achievable, with existing patents and products showing that interchangeable footwear components are feasible. However, due to the distinct biomechanical demands of sneakers versus cleats, especially in high-performance sports, the design must be carefully optimized. Focusing on one sport (like soccer), prioritizing interchangeable outsoles, and developing a secure, user-friendly locking mechanism are key recommendations. Success will also depend on using advanced materials and conducting thorough safety and biomechanical testing. With innovative engineering, such a product could appeal to athletes needing quick transitions between surfaces or recreational users seeking versatile, cost-effective footwear.

Based on the recommendations, we refined the concept to focus specifically on soccer cleats with interchangeable outsoles. Very different from retractable cleats, but intriguing. Also, weird, which seemed useful.

Now, with the pieces in place, it was time to put AI to work on visualization. We do this the human way all the time at Spark. Could AI help us push the boundaries even further? Let’s take a look.

After testing three large language models (LLMs) and iterating on the prompt a dozen times, we ultimately landed on Perplexity. Here’s the final prompt we used:

“Design a futuristic pair of soccer cleats with interchangeable outsoles. The cleats should feature a visible, high-tech interchangeable sole—function meets aesthetic. Use a color palette of acid green, toxic pink, and infrared orange—retina-burning, high-contrast hues that make the product feel aggressive and bold. The overall design should lean into a neo-futurist aesthetic: glitchy yet sleek, with an almost alien quality. Imagine if H.R. Giger and a cyberpunk industrial designer collaborated on athletic footwear—mechanical, sculptural, and a little unsettling.”

So how did AI do?:

We-e-lll, here’s where AI ran into a little trouble.

After many iterations and prompts, it struggled to capture the brand’s emotion and missed the mark on the product’s unique features—the interchangeable sole.

Because AI image generators rely on existing visual references, they had trouble accurately rendering the cleat’s mechanism and distinctive style. No matter how we approached it, we couldn’t land on an image that worked both with and without the cleat. As it stands, the visual is pretty confusing and a bit off—why are the cleats facing upward? It just doesn’t quite make sense.

In the end, it took human intervention to guide the concept and shape the final design.

We built a version. AI built a version. Then we combined them.

A solution that’s both more compelling and more coherent than either approach on its own:

At Spark No. 9, this is the kind of process we live for: taking a half-formed idea and pushing it through strategic thinking, technical validation, brand exploration, and vivid product visualization. With AI as our assistant, we’re not just dreaming up what’s next; we can actually see it, test it, and understand how it might function in the real world. Twisted Pair is just one example of how imagination, strategy, and the right tools can turn blue-sky concepts into grounded, market-ready innovations.

Needless to say, we’re not ready to go to market with this concept. However, the exercise raises an important question: are we close to a world in which anyone can develop a new product? Are we about to see scads of tiny brands chipping away at the dominance of big players? How is the competitive advantage of big brands changing in an AI world?