Generative AI is Disrupting Product Development. Here’s How.
Today I had the opportunity to participate on a panel discussing how Generative AI will impact product development. It’s a topic that would have felt far off in the future a year ago, but in the last 6 months the explosion of the trend has opened the door to new opportunities, and many are left wondering how it’ll fit into their strategy.
Like so many, we’ve been using Artificial Intelligence for years to improve internal processes and customer experiences, but the use was often behind the scenes.
Now, Generative AI is front-and-center, placing creative power in the hands of millions, poised to disrupt the way we work and build solutions.
It’s 1997 All Over Again
I want you to transport yourself back in time. You’re a marketing executive at a retail company in 1997, and your team just told you they need to photograph every product you have from multiple angles and put them on your soon-to-be-released website. What a mountain that looked like.
And now, we’re in a similar stage of disruption for product listings, today.
Generative AI presents a very real opportunity to explore the personalities of your products. For example, a user might search for “road trip destinations” today, giving you some clear guidance on the types of keywords you need to use to be found, and the places you might need to be listed to be discovered.
But what happens when the search becomes more like a prompt:
“I’m a 40-year old dad of two boys, 10 and 7. We want to take a road trip from new York to somewhere that has a great beach with a boardwalk we can have dinner on. We’re happy to drive two days, but never more than 6 hours in a car to get there. My kids need a separate room from my wife and I, and one of the boys is sensitive to gluten so that will impact where we can eat. Plan the trip, meals, activities for the kids, and itinerary, giving us at least 2 days of relaxing on the beach.”
That could change how you list your products. It’s not about showcasing your offerings, it’s about creating a deep understanding of the way your products will be used, the feelings they might evoke, and the context in which they’ll be described.
It might also give you unique opportunities to build products that fit specific emotions, feelings, or desires. With search phrases giving way to descriptive prompts, the data of what a consumer is looking for could empower hyper-niche products to accelerate their growth.
Community Disrupted
Generative AI is going to spark a lot of questions about where product information comes form, and how the aggregators of that information add value to ensure they’re still useful.
In the trip planning example, there’s a good chance that the LLM was trained on data from Yelp, TripAdvisor, Expedia, or a myriad of other solutions who have built their brand on the idea that they would deliver the right answer from search results.
We’re witnessing some initial exploration of this shift within solutions like ChatGPT, as brands build widgets to integrate their information more natively. The question remains how well that’s adopted — for all the “skills” enabled by companies for voice products, most people still use the most basic native functions of Alexa and Google Home.
Prompting the Product
The past decade has seen an explosion of boutique brands, creator marketplaces, and hyper-niche products.
Prompted products could become a very real way for more creators and companies to further their journey of making timely, relevant products with significantly less cost.
Imagine that tomorrow, a hit show has its final finale. Let’s call it, Ted Lasso 😉.
Fan art for the show is already abundant, and while many creators are designing from scratch, and ever-growing number are using templates in sites like Canva. With generative AI as a helping hand, products could be created that reference fan-favorite moments from the finale and listed on creator marketplaces before the show is over.
The ability to make variations of the fan art to fit different complementing styles of clothing or aesthetics will be easy to do with only a handful of additional prompts.
For brands, creators, and marketplaces, there will be questions about how consumers react. Does the expanded variety result in decision fatigue? Does the artist or brand develop their own unique style based on the types of prompts they write, akin to the brush strokes of a classic artist?
Or do consumers instead become the creators themselves, writing the prompt, gathering the output, and printing the item using an online on-demand shop, all in a matter of minutes?
Welcoming A New Member of Your Dev Team
A meme has been circulating the techno-sphere for a few months now. The headline says “For AI to replace designers and developers, product owners would need to accurately describe what they want. We’re safe”.
As a technical product owner, the joke isn’t lost on me.
But the real opportunity is not in replacing anything, it’s adding a member to your dev team who can help improve the efficiency of the non-additive work.
Let’s look at this from two perspectives: the product owner, and the designer.
The designer has some amazing opportunities to make their least-fulfilling work more efficient. When their product owner asks for a new page to be designed for a feature that’s in their backlog, generative AI might become a supplement to the component library used today. With a few sentences, the designer prompts the AI to get them the first draft of the design, perhaps based on existing pages on the site or app, organization design guidelines, component libraries, and some basic understanding of what the new functionality needs to perform. Once generated, the designer flips through their options, grabs the one that most closely resembles their vision, and cleans it up to fit their exact needs.
The designer has been freed up to work on their more fulfilling work, including deep thinking, customer experience research, and net new builds that require unique human skills.
The product owner, after receiving the design from their teammate, no longer needs to go back and forth on revisions for minor changes. “Our development team say’s the button needs to be 10 pixels smaller, while remaining ADA compliant and aligned to our design standards” might be a prompt that cuts out days of waiting to prioritize work, resulting in faster speed to deployment.
The Fear of the Unknown
I want to leave everyone with some thoughts on the question what was on everyone’s mind — how will this impact my career?
Some reflection on the past can give us perspective on where the world is going.
I once heard an economist say “We wouldn’t have podcasters if we still had to farm”. It’s a bold statement, but it’s also likely true. The agricultural revolution empowered people to find work off of a farm. Since basic needs were met with fewer hands, we could focus on other work, spurring the industrial revolution that followed and the boom of manufacturing.
We have to wonder if the data-centric world we live in would be as advanced today if not for Microsoft Excel. A program of my size would have had a bookkeeper in the 1970’s, but not today. I run my own budget without any intervention from a bookkeeper, but people with a love of finance and mathematics didn’t disappear. The advent of Excel freed their minds to even more meaningful work in data science, engineering, modeling, and quantitative computing.
Generative AI feels scary for so many because it’s the first time we’re seeing AI come for knowledge work. Just this week, I watched as Adobe rolled out “Generative Fill”, and my heart sank as I thought of how long I spent practicing cloning and airbrushing in my dorm room while taking “Introduction to Graphic Design”.
But the sinking feeling didn’t last long…it was immediately met with joy for how this will bring about new possibilities we can’t even imagine yet. And that’s the beauty of innovation.
More information and a recording of the Tech Talk referenced in this article can be viewed at method.com/generativeai.
