In 2023, we’re seeing the rapid roll out of AI tools. It has ignited a massive paradigm shift in computing, and in turn the tech industry. An AI arms race is happening between OpenAI, Google, and soon X.AI (Elon’s venture into AI). In response, companies are quickly scrambling to add AI functionality to their products, and entire product roadmaps are being redefined across the industry.
Unlike the hype cycles of the past few years, this wave feels different. This feels more like a technological revolution akin to the invention of the printing press. While history doesn’t repeat, it does rhyme – the printing press made words cheaper and more widely available, and the roll-out transformed the way information was produced and disseminated. It made knowledge accessible to the masses and lead to the growth of various industries and administrative systems. It seems these AI tools will do the same at a way bigger, and faster scale.
Like many of us, I’m both excited and anxious. I’m enjoying playing with these new tools, but I’m also observing the rapid proliferation of tools and with each advancement, I’m thinking more about their impact on me as a product designer. It’s early and hard to say, so I’ve been in a state of play and signal collection mode for the last few months (which feels like the best place to be right now for sense making and adaptability).
In my sense making, I’ve run some thought experiments and on two key angles –
- What it means for software design
- What it means for the product design role
Here are some speculations…
- For software design
There is a massive paradigm shift in computing underway. In some cases, entire apps and experiences will need to be rethought and rebuilt from the ground up.
In the last decade+, we saw the proliferation of design systems and imperative UI — interfaces that require a person to provide detailed instructions for a computer to complete a task. The result was a lot of interfaces full of form fields and inputs.
As we enter this new era of computing, we’re seeing the rise of declarative UI — interfaces that allow a person to specify a desired outcome, leaving the rest to a system. The initial implementation of this has been Natural Language Interfaces (NLIs) in the forms of chatbots.
Put another way, we’re iterating on human-computer interaction and making it easier for humans to get a computer to do what they want with less input, and the computers are understanding the way humans speak, rather than humans having to learn how a computer speaks.
With this being true, this means entire interfaces could be replaced by a dialogue box or some other interaction pattern (e.g. voice) that enables human conversation. And as we consolidate interface components into less, perhaps designing digital experiences starts to focus more on designing personality and tone than layout and hierarchy. It’s hard to say. Sure, there will always be a need for conventional UI that explains a system to a human, but I’d posit a lot less of it in many cases.
2. For the product design role
With shifts in how we design software also comes shifts in how we work. The core of the product design role still remains (understanding people, solving their problems, and using communication to get design decisions shipped to the world), but the means for which we achieve these has changed.
Understanding people remains a human task. And while nothing substitutes human to human interaction for gaining insights, it is true that AI is already able to provide a sense of human perspective on different topics with the right prompts. For example, it can be used to brainstorm human pains, gains, and jobs to validate through research, or it can be fed insights and help spot themes and extract insights. We can only assume with more connections to data sets and real time information, AI will get better and better at understanding human beings.
Solving problems is super charged by AI tooling. With an understanding of human needs, product designers engage in a divergent process of ideation and creation. They work through many possibilities before converging on the best solution to solve a given problem. AI reduces the toil of solving problems at every step of the way — even in it’s nascent form. For example, AI can come up with job stories, prioritization matrixes, interface concepts, and more with the right prompts. All of the artifacts and methods used in the design process (which were already a means to an end) can be outsourced and automated, unlocking efficiencies in design. For now these tasks need to be initiated by a human with clever prompts, but eventually I’d posit AI will develop its own methods and artifacts for the design process.
Using communication to get design decisions shipped to the world — one of the most important aspects of the product design job, is also enhanced with AI tooling. Designers use visual and written communication to articulate design decisions and get alignment with teams and stakeholders. We demo prototypes and annotate design files so that our teams can see the proposed solution to a human problem. AI can already help with the some aspects of this — through annotations, challenging design rationale, structuring demo narrative, and more, and there’s already rumblings of AI-driven UI design tooling.
If AI can already support in so much of the core tasks of the role, then what does the future of product design look like as tools advance? Who knows. But in the same way we need to start designing declarative UIs for outcome based interaction, I think designers need to focus more on outcomes than process.
It was already true that outcomes were more important than artifacts and methods in the design process, but it is even more true now. That is to say, if everyone has access to AI that can act as an assistant and speed up the design process, then the outcomes (not outputs) of design are most important.
What does this all of this mean in practice for product designers? Again, it’s early, but I think it means…
- Methods and artifacts are a means to an end. They are outputs. Outputs ≠ outcomes. A designer’s effort needs to be focused on speed, execution, and alignment towards a desired outcome. Outputs should be outsourced where possible to gain efficiencies in getting to outcomes.
- Writing becomes one of the most powerful tools in the product design toolbox. To write is to think. Writing helps us think deeply and solve problems, and it also happens to be the main input for AI prompting. More than ever, designers need to level up their ability to write clearly and effectively.
- It’s important to get hands on with AI and create more. This is the golden age of creation — in the short term we’ve all gained huge productivity leaps if we prompt effectively. To get better at prompting, we need to play with these tools and learn how to use them properly to achieve desired outcomes.
- It’s important to become an expert on people. Designers should be laser focused on a deep understanding of the people they’re designing for. Time saved on outputs can and should be reallocated into getting closer to the people that use a product.
- Systems thinking is incredibly important. Currently, AI is good at executing on tasks with effective prompts. But like an intern at your side, it needs guidance to get to the right outputs, and those outputs are part of a bigger process directed by you (the human). Unless it has access to all the relevant information of a system, AI cannot yet think at higher level abstractions and piece together these outputs towards a bigger aim. That is up to us. So, product designers should focus on levelling up in systems thinking and developing more holistic understandings of the systems they’re working on, so that they can effectively spot problems and leverage AI to support in solving them.
- Design needs to speak business. When design outcomes become more important, so does translating them into business outcomes. Designers don’t need an MBA, but a basic understanding of commerce and business acumen is necessary for articulating the business case around the design ideas we pitch. With basic fluency in business, we can prompt AI and leverage it to do the heavy lifting in building more robust business models and cases.
- Build more. A reduction in the time it takes to generate design outputs means we have more time to build real things. The “No Code” movement combined with AI provides a plethora of tools that can speed up the time it takes design decisions to become real world solutions. Designers should be getting comfortable with a basic understanding of code folder/file structures, and using no code tools and AI to shorten the time it takes to ship design outcomes to the world.
- Overall, specialize in generalization. Those who are the most adaptable are the most likely to succeed in this technological revolution. There is too much risk in being hyper-specialized in a specific design skill these days when AI can specialize rapidly and output quicker. Design generalists are the most well suited to evolve and succeed in this new age of AI.
