If you’ve ever tried to design your own pair of shoes or, frankly, any other product, chances are you’ve run into the same problem a lot of us have: you’re not a designer.
Laypeople often choose too many colors, jump into the process without a plan, and, basically, create something that’s—to put it lightly—less than ideal. It doesn’t take too many attempts to realize that design is a lot harder than it looks.
Designers themselves have different problems. They don’t need much help actually designing products—it’s their jobs, after all—but forecasting trends can be tricky. No matter how many consultants they employ and how much market research they do, chances are, their customers are sending signals nobody’s picking up.
The question we’re going to answer today is if artificial intelligence can do anything about these challenges. Can AI help in the design process? How good is AI at being creative? And, if it can be creative, is it going to take designers’ jobs?
We’ll answer the last one first: no. Artificial intelligence isn’t a job-stealing boogieman. Most often, what AI does—and will do in the design space—is give people an extra tool to use that makes them more effective. It’s a technology that helps people work smarter, not replace them.
A good analogy here is something that actually isn’t AI: the spellcheck on your word processing program. Before these existed, before you saw a squiggly red line underneath your typos, editors had to catch each one. Did spellcheck replace them? Nope. Instead, it gave both authors and editors a bit of help. It made writing easier and editing faster.
AI will function in the design space in a similar way. But how exactly? Well, first, we need to understand what AI can do in the space.
Artificial intelligence—specifically deep-learning neural nets—are a lot further along analyzing images and video than most people realize. AI can look at a product and analyze it over hundreds of vectors, mapping and understanding the contours of the product, the color, the describable traits and even the indescribable ones. And, while humans can do this too, where AI excels is that it can understand vast numbers of products in far less time than a person can. It can learn an entire catalog—hundreds of thousands, even millions of SKUs—in a way that a person simply can’t.
Now, what we do with that information is where things get interesting.
Consider that an AI could understand a giant shoe brand’s entire catalog since the 60s. How could a designer use this? She could make sure new designs she was creating fit into the brand’s overall philosophy. She could see where the design slots in, what it’s similar to and what it isn’t. This might help her avoid essentially recreating a dud or accidentally plagiarizing an old design. If the AI is trained on a vast enough space, it can make sure she doesn’t plagiarize a design from a competitor too.
AI can also help non-designers create bespoke products. Brands like Adidas and Reebok have certain shoes you can essentially design yourself. You’re given a popular shell, a baseline shoe, and then you choose the colors and fabrics of each part, from the instep to the toe to the tongue to the laces. AI can help make sure people don’t create something garish by gently guiding them towards common color palettes. For that matter, it can help designers do that too–not that they need the help, but, like the spellcheck functionality we mentioned above, helping designers slot their product into existing brand paradigms can be important.
Probably the most powerful way AI can help designers though is trendspotting. Because AI can understand both entire catalogs and how customers interact with those catalogs, it can start to forecast which products or styles are slowly gaining popularity. It can look at customers who view certain items and what else they like so designers can create products for particular audiences. It can, by aggregating millions of clicks and customer actions, identify the similarities between sought-after products so that designers can stay on top of what their customers actually want.
In the future, we might even see it combining popular trends and trying its hand at design itself, though, if AI’s attempts at music and fine art are any indication, these products will probably be a little flat.
The point is: AI is coming to the design process. It can help design products by providing reasonable guard rails that can be transcended by great designers or followed by laypeople. It can help spot trends so that designers can predict the next big thing.
And really, those are just two examples. Because AI can understand vast product spaces and compare customer actions as they interact with those products online, designers will have access to perhaps the perfect dataset to help them excel. How they use that to create more exciting products is up to them, but, make no mistake: AI will be coming to the design space.
It’s just a matter of who uses it best.
Andy Narayanan is the VP Intelligent Commerce at Sentient Technologies.