AI Design Culture | Design Portfolio

The Philosophy

A strategic manifesto defining how I bridge the gap between human creativity and artificial intelligence. "AI is not a magic box; it is an engine for augmentation." My approach to AI leadership moves beyond simple tool adoption to building an AI-Empowered Design Culture. I view AI not as a replacement for the designer, but as a force multiplier that demands a shift in how we value design work.

The "Human-in-the-Loop" Protocol: I established a workflow where AI serves as a rapid generator, but the designer remains the strategic decision-maker. We treat AI outputs as drafts, not deliverables—enforcing a protocol where every artifact is verified by a human expert to ensure it aligns with user context and ethical standards before it reaches engineering.

Managing the "Magic Box" Expectations: A core part of my leadership involves cross-functional education, correcting misconceptions from Product managers that AI can "just do the design." I bridge the gap between design intent and technical feasibility by explaining concepts like Contextual knowledge and user needs vs business needs to manage stakeholder expectations standards before it reaches engineering.

Automate the routine : I actively guide teams to automate the tedious tasks like reviewing documentation for information retrieval and routine UI construction to free up mental bandwidth for the "critical" tasks that need a human input like user empathy, complex problem solving, and strategy.

The "NPDS Guru" (Custom Agent powered by RAG)

NPDS stands for Nelnet Product Design System. Solving the "Design System Bottleneck" through proprietary AI architecture.

The Challenge

The impact

Operational Efficiency

We achieved a 67% reduction in chat queries and support meetings for the Design System team, allowing them to focus on system evolution.

Following a major shift to a new Design System, our design team faced a "knowledge bottleneck." The documentation was comprehensive but dense, and distributed across too many files. Designers were struggling to find answers quickly, which led to a high volume of repetitive support queries and increasing "design debt" due to inconsistent component usage

The agent successfully democratized access to complex system rules, reducing the "time-to-answer" from minutes to seconds.

67%

The Strategic Solution

I realized we needed an interface that could bridge the gap between the designer and our documentation. I rejected the use of a generic public LLM due to data privacy concerns and the risk of "hallucinations." Instead, I conceptualized and deployed NPDS Guru, a custom agent powered by RAG (Retrieval-Augmented Generation)

45%

Quality Control

We saw a 45% reduction in design debt tickets because the "correct" usage rules were instantly accessible

90%

Adoption

By positioning the tool as a workflow accelerator rather than a replacement, I secured team-wide buy-in and turned a source of friction into a competitive advantage

more explorations coming soon...........

My work in bridging the gap between design intent and technical feasibility is ongoing. Stay tuned for upcoming deep dives into Multimodal AI workflows and Agentic interfaces...

Building an AI-Empowered Design Culture