Nasir Uddin on Avoiding ‘Average UI’: The Right Way to Integrate AI in Design

In a recent LinkedIn post, Nasir Uddin discusses a common pitfall many designers encounter when using Artificial Intelligence (AI) for User Interface (UI) design, arguing that subpar results often stem from an incorrect workflow rather than the AI itself.

Nasir Uddin, drawing from his experience with hundreds of designers at Musemind, observes a recurring pattern of designers jumping directly to AI prompts without foundational steps. “Jump straight to prompts, skip references, expect magic from vague inputs,” Uddin writes, characterizing this approach not as design, but as “guessing.”

The Flaw in the Standard AI Design Workflow

According to Nasir Uddin, the prevalent method of using AI for UI design, which often involves immediate prompting and a lack of proper grounding, leads to what he terms “average” results. He posits that AI is a powerful tool, but its effectiveness is contingent on how it’s wielded. “Not because AI is bad. Because the workflow is wrong,” Uddin emphasizes.

Uddin’s Five-Step System for Superior AI-Assisted Design

To achieve better UI outcomes with AI, Nasir Uddin proposes a fundamental shift in the design process, detailed in a carousel shared on LinkedIn. This system emphasizes that AI should augment, not replace, core design thinking.

Step 1: Prioritize Real References

Nasir Uddin stresses the importance of starting with tangible examples. “Great UI never starts from nothing. It starts with real products and real patterns. Not Dribbble concepts. Actual shipped UI,” he states. This grounding in existing, successful designs provides a solid foundation before engaging with AI tools.

Step 2: Establish Direction Before Prompting

Before generating any AI prompts, Uddin advocates for creating mood boards and defining the aesthetic and emotional tone. “Define the style. Define the tone. Define the vibe,” he advises, suggesting the use of platforms like Pinterest or analysis of live applications to set this direction.

Step 3: Provide Clear Context to AI

Vague inputs yield poor outputs, according to Uddin. He recommends giving AI specific instructions, such as analyzing spacing, typography, color palettes, and existing components. “Clear input creates better output,” Uddin notes, highlighting the direct correlation between prompt specificity and result quality.

Step 4: Build Design Systems First

Nasir Uddin argues for a top-down approach where design systems are established before individual screens are designed. This systematic method, he explains, is where AI can offer significant time savings. “This is where AI saves real time,” he writes.

Step 5: Structure for Scalability

The final step involves refining AI outputs into production-ready assets. This includes converting outputs into variables and creating proper components, ensuring the design is scalable and maintainable. “Turn outputs into variables. Create proper components. That’s how it becomes production ready,” Uddin explains.

AI as an Accelerator for Strategic Designers

In conclusion, Nasir Uddin reiterates that the same AI tools can produce vastly different results based on the designer’s approach. He differentiates between random prompting, which leads to average UI, and a reference-driven, systematic process that yields premium results. “AI is not replacing designers. It accelerates designers who think in systems,” Uddin asserts, prompting readers to consider whether they are truly designing or merely prompting.