AI & the future of product design (predictions)
As a digital product designer, a lot of what I do could be (and should be) automated. Tasks like briefs, engineering hand-offs, and QA could all be streamlined with AI. However, the quality of the output depends heavily on the quality of the input. That’s why, in my opinion, we’ll still need people with a trained eye along with deep subject matter expertise to guide and train AI models to do specific tasks.
Currently, tools such as Chat GPT and technologies like Stable Diffusion (Mid journey/Dalle) are primarily productivity tools, with potential for future developments. So let’s assume that we will all still have our jobs this time next year and focus on how we as designers can maximise the use of these tools to improve our workflows and increase efficiency — unlocking more time for collaboration, problem-solving and creativity.
1. Insight Gathering
Insight gathering is a critical phase in the product design process, as it provides the necessary understanding of user needs and pain points, trends, and behaviours from both a qualitative and quantitative perspective. These Insights then go on to influence product roadmaps, prioritise initiatives, and affect the way we think about problem solving and eventually defining solutions. The biggest opportunity I see with AI in this area is the ability to impart unbiased insights in real time. Whether it’s through the use of predictive analytics or some other means, AI could identify patterns in user behaviour, make recommendations based on data, and streamline the time it currently takes to gather and interpret insights by 10 fold, helping designers make better, more informed decisions as they design.
2. Automated Design Tasks
Automation has already impacted various parts of the design workflow, from automated resizing and file exports to AI-assisted colour palettes and font suggestions. With the advancements in AI technology, we can expect even more tasks to be automated in the future, such as brief creation (ie isolating specific problems to solve related to goals); estimating project scope and timelines more accurately based on complexity and productivity; generating starter kits complete with an automatic assignment of design systems and flows, while simultaneously setting up a project with engineering which contains all the parameters and boiler-plates required to streamline development; generating presentation slides that build themselves as you work through the solutions in preparation for stakeholder reviews… basically reducing repetition and improving efficiencies at every twist and turn throughout the design process, freeing up designers to focus on more complex design challenges, creativity, and problem-solving.
3. Design systems
Design systems have become an increasingly popular approach to streamline the design process, ensure consistency, and speed up production. I’ve been a huge fan since the inception of Airbnb’s DLS. But with all the talk about design systems, they are inherently flawed if 1. they’re not maintained, 2. they’re not codified, 3. There’s no documentation, and 4. the design system is out of sync with production. With AI, the potential to create and maintain design systems could be enhanced even further. AI could analyse the usage patterns of design system components, identify inconsistencies, and recommend updates or new components based on user needs, product and team requirements, or shifts in market trends. Additionally, AI could assist with creating and managing design tokens, automating the conversion of designs to code, generating design documentation on the fly, and updating production automatically to ensure everything is in sync. Magic! 🪄✨
4. Use-case (flow) generation
Use-case (flow) generation is another area where AI can help improve the design process. As designers, we often spend a lot of time mapping out user flows, wireframes, and prototypes to ensure a seamless experience for the end user. However, this can be a time-consuming and iterative process, with multiple rounds of feedback and revisions. AI can assist with this by analysing user behaviour and generating potential use cases and flows automatically based on conventional wisdom, ease of use, and limited friction. By leveraging AI to generate commonly used, tried and tested use-cases, designers can focus on refining and iterating on the most promising or unique flows rather than starting from scratch each time. For example, there’s no need to redesign a consumer login flow every time you create a new product, or for that matter a payment flow, a booking flow, or a customer feedback flow.
5. Content and tone
Content and tone play a crucial role in communicating the brand message and values to the user. Traditionally, this is a task for copywriters and content strategists, who work closely with designers to ensure the overall user experience aligns with the brand vision. However, with the help of AI, we could see significant improvements in the content and tone of digital products. Natural Language Processing (NLP) and other AI-powered tools can analyse user feedback and engagement data to determine the most effective language and tone to use in different contexts. AI could also be trained on a company’s brand voice and messaging guidelines, allowing designers and content creators to maintain a consistent tone throughout the product.
6. Collaboration
Collaboration is a critical component of any design process, and it’s no different when it comes to incorporating AI. As AI becomes more prevalent in design workflows, collaboration will become increasingly important between designers and engineers, as well as between designers and AI. Designers will need to work closely with AI systems to ensure that the output aligns with the design intent and meets the project’s goals. Additionally, designers will need to collaborate with engineers to ensure that any AI-generated designs are feasible and can be implemented within the product’s technical constraints.
AI Co-creation (COAi™) is an exciting concept, and we are already seeing some promising Figma plugins that leverage GPT4+ technology to generate UI designs based on prompts. For example, the “DesignLingo” plugin generates design terminology to help designers write more effective design briefs. “Magician” from the folks at Diagram generates icons, images, and copy fromm text prompts, and the “Design Me” plugin generates design concepts based on keywords and adjectives, providing designers with a starting point for their designs.
In the future, it’s not hard to imagine a world where we input a detailed brief (2), and the AI generates entire design flows (4) containing up-to-date components (3). This would make the role of a designer more like a Design Director, co-creating with AI to achieve the desired result. However, one of the biggest questions in my mind is whether someone with no design background or experience could co-create designs at the same level as someone who does, with AI. Would it require a “trained eye” or not? 🤔
7. Design Review & approvals
I’ve experienced both sides of the design approval process — presenting my work to review panels, stakeholders, and CEOs, as well as giving feedback to other designers. While design reviews have many benefits, they often create bottlenecks in the product and engineering workflow, causing frustration for product managers who prioritise speed and MVP’s over everything else. Designers have to navigate the tension between advocating for users, high-quality solutions, and systems, while also satisfying the demands of their product teams.
AI could provide a solution to this challenge — codifying design principles, removing subjectivity from feedback, and aligning stakeholders well in advance of the design review, and in some cases eliminating the need for a review altogether!
8. Eng hand-off & QA
Given all the advances in design tools and systems over the years, I still see designers “red-line” their work in preparation for engineering. Despite all the red-lining, engineers still fail to replicate designs accurately, leading to quality issues, design debt, and, worst of all, unhappy designers! So why, in this day and age, haven’t we evolved beyond red-lining, and how can AI play a role?
If we think about a mature startup and product team made up of the usual combination of engineers, PMs, and product designers, there should be some predictable workflows and methods applied to building and shipping new products and features. In this world, an AI could act as a “quality assurance guide” across the entire production process. In fact, engineers already have AI systems in place when it comes to coding and code reviews, so training the AI on design systems to interpret mocks and recommend components, design patterns, and interactions, or even construct the scaffolding of a single view or use-flow, doesn’t seem too far-fetched.
A Quality Assurance AI (QAAI™) could save thousands of hours in reviews, reduce the amount of errors and bugs being shipped into the wild, improve engineering time (PMs would love this), and overall improve the end user experience! Win win.
9. Data Analysis
Data analysis may fall a little outside of the conventional product design workflow, but it does have a meaningful effect on insight gathering (1). Currently, we devise product roadmaps using a combination of data and UX research, with data playing a more significant role in my experience. Analysing core metrics is a tedious job that involves analysts and product managers rifling through dashboards, extracting key insights from the latest experiments and production builds to then report back to the product team.
What if AI, which is trained on the inner workings of your product, could provide feedback in real-time, along with recommendations, making the analysis process more inclusive for non-technical staff and enabling the entire product team to problem-solve on the fly?
10. Values and Principles
Quite often, we approach design problems based on personal beliefs, values, principles, or a collective mission. In the corporate world, this plays out at both the macro level (the entity) and micro level (the individual). Ideally, you want both to align to avoid dissent and complications between conflicting values (although I strongly believe that a little dissent goes a long way when it comes to challenging groupthink).
At the entity level, a company will more than likely define a mission that acts as a north star vision (e.g., “Belong Anywhere”), along with a set of core values and principles (e.g., “Be a host”, “Embrace the adventure”, or more notoriously, “Move fast and break things”). These function as a rallying cry as well as a touchstone to help with decision-making when trade-offs need to be made or behaviours called into question.
What if there was an entity-level AI trained on these specific details of the company mission and its values, which could be deployed (or be ever-present) to guide a company closer to its goals and achieve its stated mission?
How would this affect the way we design or what we’d design?
While the idea of AI potentially taking over our jobs can be a little nerve-wracking, there’s no denying that AI has the potential to revolutionise the design process by reducing repetition, improving efficiencies, and allowing more time for problem-solving and creativity. Some of the biggest opportunities and changes will come in the form of unbiased real-time insight gathering, automating repetitive design tasks, generating potential solutions, enhancing design systems, and reducing the friction between design and engineering.
As a designer that’s ridden the many waves of technology over the past 25 years, these changes have been a long time coming. But one thing remains true — a designer’s superpower lies in their ability to harness technology to its fullest potential, resulting in better and more impactful products and user experiences. The same is true for AI.
If you have any other predictions on how AI could affect product design, feel free to reach out. I’d love to collaborate!
