AI-Assisted Product Discovery
From product requirements to functional prototypes in two days using AI-supported workflows.
Impact snapshot
Faster concept validation
Functional prototypes allowed product managers and engineering leads to evaluate interaction flows earlier in the discovery phase compared to traditional wireframes.
Reduced discovery time
Two product features were explored and prototyped within two days, significantly shortening the time between product requirements and interactive validation.
Improved collaboration
Interactive prototypes helped facilitate more productive discussions between design, product, and engineering teams by providing realistic interaction scenarios.
Earlier stakeholder alignment
Client validation occurred earlier in the process, reducing uncertainty around feature direction before moving into full UI design.
Repeatable discovery workflow
This experiment helped define a practical approach for integrating AI-assisted prototyping into early product discovery while maintaining structured UX practices.
AI-Assisted Discovery
Exploring feature concepts using product documentation and AI-supported exploration tools.
The discovery phase started with existing product documentation, including Confluence PRDs and initial feature tickets. Instead of relying solely on traditional wireframing, I experimented with AI-assisted tools to accelerate early concept exploration.
Using tools such as Codex, Google AI Studio, Claude, Cursor, and GitHub Codespaces, I explored different approaches to generating interface structures and interaction logic directly from the product requirements.
The goal of this phase was to quickly explore possible layout directions, interaction patterns, and feature structures that could later be refined through established product design practices.
Rapid Prototyping & Internal Iteration
Transforming early explorations into functional prototypes for product and engineering discussions.
Based on the initial explorations, I structured the generated outputs into clearer layouts and interaction flows using my experience in UX methodologies and technical constraints.
Functional prototypes were created and deployed using platforms such as Vercel and GitHub Pages, allowing stakeholders to interact with early versions of the features directly in the browser.
These prototypes helped product managers and engineering leads review the concepts in a more realistic context, making it easier to discuss feature feasibility, interaction logic, and potential edge cases.
Stakeholder Validation
Aligning product direction through interactive prototypes and early design discussions.
Once the internal iterations stabilized the core flows and layouts, the concepts were shared with the client for validation.
The interactive prototypes allowed stakeholders to experience the proposed feature structure and interactions directly, providing more clarity than static wireframes or documentation.
Through these discussions, the proposed layout and interaction flows were validated, confirming alignment between product goals, user experience considerations, and technical feasibility.
AI-Assisted Prototype
Generating functional prototypes directly from product documentation using AI-supported tools.
During the discovery phase, I experimented with AI tools to rapidly generate functional prototypes based on existing product documentation, including PRDs and initial feature tickets.
Using tools such as Codex, Google AI Studio, Claude, Cursor, and GitHub Codespaces, I explored different approaches to translating requirements into working interface structures and interaction logic.
The goal of this step was not to create final design solutions, but to quickly produce interactive prototypes that could help validate feature structure, layout direction, and interaction flows.
These prototypes were deployed using platforms such as Vercel and GitHub Pages, allowing internal stakeholders to interact with the early concepts directly in the browser.
From Prototype to Product UI
Refining AI-generated exploration into structured UI using UX methods and the design system.
While AI tools helped accelerate the creation of early prototypes, the next step required applying structured product design practices to refine the experience.
Using my experience in UX methodologies, interaction design, and knowledge of the product’s technical constraints, I refined the layouts, clarified the interaction flows, and aligned the feature structure with the existing product architecture.
The validated concepts were then translated into production-ready UI screens in Figma using the existing design system components, ensuring consistency with established patterns and interaction guidelines.
This step bridged the gap between rapid AI-assisted experimentation and structured product design delivery.
AI Tools & Prototyping Environment
Experimenting with multiple AI and development tools to explore the most effective workflow.
During this exploration, I tested several AI-assisted tools and development environments to understand how they could support rapid product discovery and prototyping.
Tools used during the process included Codex, Google AI Studio, Claude, Cursor, and GitHub Codespaces. Functional prototypes were deployed using platforms such as Vercel and GitHub Pages to enable real interaction testing with stakeholders.
This combination of AI tools, development environments, and product design practices allowed for faster exploration while maintaining alignment with technical constraints and product architecture.
Traditional vs AI-Assisted Workflow
Using AI tools to accelerate exploration while maintaining structured product design practices.
Traditional product discovery workflows typically rely on wireframes and static prototypes before functional validation. By introducing AI-assisted exploration and rapid prototyping, it became possible to validate interaction flows and layout structures earlier in the process.
Rather than replacing product design methodology, AI tools acted as an acceleration layer for exploration, enabling faster iteration cycles and earlier collaboration between design, product, and engineering teams.