Embedding AI Across the Full Product Development Lifecycle.

 

Challenge  

UX workflows at key friction points - stakeholder alignment, research synthesis, content quality, and rapid prototyping - relied on manual effort that slowed delivery and introduced ambiguity between discovery and execution.

​

Approach

I identified five discrete workflow gaps across the product development lifecycle and designed AI-assisted solutions for each, moving from personal productivity toward systemic impact:

• Converted post-stakeholder discussion notes into interactive HTML prototypes within hours using Kiro, giving cross-functional teams a concrete visual reference point immediately after discovery sessions.

• Contributed UX-specific steering, agent, and skills files to Delta's Enterprise AI Kit - grounded in Nielsen Norman Group heuristics - filling a gap in the AI coverage alongside observability, accessibility, and other perspectives.

• Prototyped an AI prompt workflow to transform raw meal titles and ingredient lists into elevated, passenger-facing descriptions, with a pairing recommendation extension proposed for premium cabins.

• Began integrating the Figma MCP server with Kiro and Delta's Airfoil design system to produce on-brand prototypes that reflect actual product specs rather than generic HTML scaffolding.

• Designed a structured usability research synthesis workflow - running consistent-script session transcripts through AI prompts and returning output to the facilitating researcher for validation.

​

Outcomes

• Rapid prototyping workflow reduced the gap between stakeholder discussion and shared visual alignment from days to hours.

• UX heuristic files extended enterprise AI kit coverage to include UX-informed behavior guidance, available to all teams adopting Kiro.

• Meal description and wine pairing workflow in progress - concept validated, with prompt design underway to transform existing catalog data into elevated passenger-facing copy for all cabins and sommelier-style pairing suggestions for premium cabins.

• Figma MCP integration concept - if completed - would eliminate the "imagine this in our design system" disclaimer from every prototype review session.

• Research synthesis workflow reduced manual analysis time while maintaining validity through consistent scripting and facilitator review.

​

Key Learning

AI delivers the most durable value not as a standalone tool but as a precision layer applied at specific handoff points - where ambiguity accumulates, manual effort compounds, or quality degrades due to time pressure.

​

Growth Opportunity

The prototyping workflow produced valuable alignment artifacts but relied on individual initiative rather than a shared team practice. A natural next step would be formalizing the workflow — prompt templates, a capture format, output conventions — so it scales beyond one practitioner and becomes a repeatable team asset.