The future of design: Generative AI turns designers into supervisors

Lauri Kieksi Strategic Digital Designer & People Lead, Solita

Published 29 May 2024

Reading time 6 min

In today’s whirlwind of Generative AI disruption, designers’ work is changing faster than you can say “pixel perfect.” There’s less and less time to find themes in that sea of post-its or to get that drop shadow effect just right. Perhaps deservedly so. Increasingly, the expectation is that designers can churn out data-driven, user-tested, business-boosting designs at breakneck speed. With Generative AI bulldozing its way onto the scene, designers might be wondering if these advanced tools are here to steal their jobs or simply lend a helping hand. Here’s our take on the role of Generative AI in design.

Driving design productivity with Generative AI

Many companies struggle to generate actual value out of the masses of data they create and hoard. Back-end systems, automation and APIs are efficient at handling and delivering structured data, but traditionally only a small slice of the data being passed around within an organization has come in a structured form fit for machines to read. While sales figures, predictive analytics or customer feedback scores pour in and are distributed across the organization through endless dashboards or automation systems, a great deal of information still remains trapped within PowerPoint decks, Excel sheets, text files, PDFs or Slack messages.

Until now, connecting the dots on unstructured data has always hinged on people. Generative AI is gradually changing this – finally letting our unstructured data become part of the organization’s fabric of ambient data instead of PowerPoints everyone forgets about.

With Generative AI, freeform data from various sources can be queried and analyzed much like the structured and machine-readable data we’ve relied on for decades. Imagine being able to query past project reports, user feedback, and concept documents as easily as you would a database of sales figures or web analytics.

Generative AI needs direction and supervision, but when done right, Generative AI can act as powerful assistants or agents for our work.

In design work, Generative AI can essentially serve as a junior assistant of everything. Research tools like MAXQDA can help uncover themes and correlations in large datasets that might otherwise go unnoticed. Large language models like Claude and ChatGPT can help create feature specifications while image generators such as DALL-E and Midjourney can help create concept visualizations, illustrations and photos. Tools such as Uizard can help generate ideas for user interface layouts while ChatGPT can even turn mockups and sketches into functional HTML prototypes.

Generative AI can act as a co-pilot throughout the design process, providing fuel for design and ways to create tangible outputs quickly, while designers remain at the helm. By taking over routine tasks and offering up ways to make sense of large amounts of data, Generative AI frees up designers to focus on the bigger picture.

Of course, human oversight remains paramount, as Generative AI models can create errors or hallucinate. Generative AI hallucinations occur when the AI produces information or content that seems realistic but is actually false or inaccurate. This happens because the AI generates responses based on patterns it has learned from training data, but it doesn’t truly understand the context or verify facts like a human would. AI models also have built-in biases depending on what that training data set is – although in many ways the same is also true for humans, and in the ideal scenario, AI and humans can spot each others’ biases and compensate.

Cumulative customer insight by DALL-E

Harnessing cumulative customer insight

Understanding people and their context is crucial for the design process. Can Generative AI create synthetic users and replace customer research and human insight? The concept of a synthetic user poses several challenges related to ethics, data reliability, and more. But there are clear achievable wins in using Generative AI to make better use of real insights.

Far too often customer insights are built in a project, only to be forgotten afterwards. Or insights are siloed in a project-specific workspace. Imagine being able to ask Gen AI to find relevant user data from various sources like past projects and have the insights synthesized into something that’s applicable to ongoing work. By tapping into this reservoir of cumulative knowledge, we can gain deeper insights, leading to smarter design decisions. This approach transforms research and insight investments from one-offs into a tool for continuous development.

We shouldn’t stop there, however. The next step is integrating cumulative customer insight directly into project management tools and design software. By integrating this cumulative insights platform into custom-built tools or industry staples like Asana, Jira, or Figma, teams can access relevant customer insights right where they work. 

Imagine a designer getting instant feedback from past user studies while creating a new interface, or a project manager pulling up user pain points when planning the next development sprint. This sort of seamless and proactive integration ensures that valuable insights are always at hand, fostering a more informed and efficient development workflow.

From AI assistant to AI teams

A lot of the power of Generative AI comes from the fact that it can act not just as a single assistant, but as a team of assistants for an individual worker, vastly expanding their capabilities. With AI tools, a designer can delve into areas that has traditionally required specialized teams of professionals. For example, AI-driven audio and video generation tools enable a designer to create compelling multimedia content without needing an audio engineer or video editor. 

Rapid prototyping tools powered by AI allow for the quick development of convincing prototypes, speeding up the ideation process. The speed of these tools opens up the the ability to test multiple high-fidelity concepts in parallel, which further accelerates the design process. The result is a more dynamic design process, where the boundaries of what a single person can achieve are significantly expanded.

While the outputs of AI tools may not always match the quality produced by teams of dedicated professionals, the ability for one generalist to orchestrate a suite of AI assistants opens up innovative ways of working. Designers can now experiment with new ideas and iterate quickly, leveraging AI to fill in gaps and enhance their productivity. This fosters a more agile and flexible workflow, where the designer can switch between tasks seamlessly, supported by AI agents handling routine and specialized tasks. Generative AI is not a silver bullet, but it is a powerful assistant that can help us unlock new levels of efficiency – but also quality.

AI as various design assistants

  • Rapid visualization of ideas: e.g.,
  • Illustration: e.g., Midjourney, Dall-e
  • Prototyping: e.g., Makereal
  • Generation of UI mockups or layouts as a basis for designer’s work: e.g., Uizard
  • Testing visual directions: e.g., Midjourney, Dall-e, Uizard
  • Eyetracking simulation: e.g., VisualEyes and Attention Insight
  • UX writing: e.g., ChatGPT and other large language models
  • Videos: e.g., Runway
  • Data analysis: e.g., ChatGPT, or specific insight tools like MAXQDA
  • Methodology assistant “How should we test this, how should we measure this…”

Best practices for using Generative AI

  • Maintain human oversight. Don’t let your critical thinking go on vacation – and always remember Gen AI needs supervision.
  • Transform customer insight nuggets into a strategic gold mine. Train AI systems on relevant, high-quality and business-specific data to turn insights into strategic assets.
  • Delegate wisely. Use AI to handle repetitive or data-intensive tasks, but leave room for human creativity, experience, and intuition.
  • Invest in training for your team to effectively use and understand AI tools to enhance efficiency without sacrificing quality.
  • Keep transparent records of how AI tools are used, including processes, settings, and outcomes.
  • Prioritize creating exceptional quality within existing budget and time constraints, rather than churning out low-cost AI nonsense at warp speed.
  1. Design