Sketch to 3D by AI

Video: 5-min pitch to the public

Duration

09/2024 to Present

Deliverable(s)

Minimum Viable Product (MVP)

My Role

UX Research
User Testing
Product Strategy

UX/UI Design

Pitching

Collaboration

Co-Founding Team
Business Analyst

ML Engineer
UX/UI Designer

Challenges

"In a world where AI moves at lightning speed, why does 3D design in the home décor industry still take 30 weeks?"

3D designers can spend months creating detailed models, yet many remain skeptical about the value of AI.

While current AI tools are fun to experiment with, they’re often deemed unreliable for professional use, leaving a significant gap in the industry’s adoption of new technology.

Roles & Goals

As the end-to-end product designer, I aimed to:

Uncover deeper insights by nudging designers to share their perceptions of AI.

Ensure timely MVP delivery through close collaboration with engineers for quality assurance.
Develop a strategic storytelling approach to drive early adoption, securing 50+ users.

Uncover deeper insights by nudging designers to share their perceptions of AI.
Ensure timely MVP delivery through close collaboration with engineers for quality assurance.
Develop a strategic storytelling approach to drive early adoption, securing 50+ users.

Research & Design

Given the testees’ (3D designers) limited knowledge of AI and reluctance to fully engage in one-on-one interviews, I implemented participatory research methods to generate honest, actionable feedback:

Think-Out-Aloud

Observed user thoughts and actions to gather preference and performance data.

Participatory Design

Involved users as co-creators to evaluate their understanding and excitement about AI concepts.

Sacrificial Tests

Used choice-based systems (like a Likert, but a harder-to-choose version) for sensitive topics

(e.g., pricing) to prompt clear decision-making.

Figure: An example of participatory research on users' perceptions of AI features

Key findings from interviewing 70+ designers in the home decor industry:

Sketch reliance

Designers heavily rely on hand sketches—refining them often takes as long as 3D modeling.

GenAI adoption

Designers expect AI to enhance creativity (e.g., mood boards, trend analysis uniquely for the home decor industry) rather than fully automate execution.

Simplified workflows

There’s a strong desire for an all-in-one tool that simplifies file transfers across different design stages, especially the fine-tuning part.

The How-Might-We Guide emerged with a central question:

How might we create an AI-assisted sketch-to-3D solution that streamlines the design process for home décor designers?

Figure: A typical user flow from sketch to 3D creation & review validated by current AI tech

Figure: 1st round of low-fidelity prototype for testing of an All-In-One key functions with 20+ users

(4 connverted to pilot users after testing)

Figure: New Mid-Fi Prototyping after first round testing & redesign

Strategic Collaboration

MVP Focus: To align with the team’s entrepreneurial vision, I proposed an iterative approach that prioritized solving the most complex industry-wide issue—fine-tuning/remeshing—while deferring non-essential features like mood boards and trend analysis.

Risk Analysis in QA: I collaborated with engineers to evaluate technical constraints and categorized tests as experimental or practical, ensuring risk-based prioritization so that we won't lost seed users.

Roadmap Leadership: I led the design strategy for cutting-edge re-meshing UX/UI experiments, balancing exploration with delivery-ready designs to keep development on track.

Figure: The latest version of experimental design - remeshing before MVP dev.

Figure: The latest version of UI before MVP Dev. - a typical design flow

Business Impact

Research Insights

The participatory UX approach doubled the amount of actionable feedback.

50%

Faster Workflows

Mid & high-fidelity prototype testing showed potential for cutting 3D creation time in half.

$20K

Funding Secured

I helped the team secure $20K in funding pre-demo by refining the tech roadmap & showcasing usability results, and pitching.

75+

Designers Waitlisted

Surpassed expectations with 50% more sign-ups, and onboarded 10+ seed users for our customer advisory panel.

Learning

Design Adaptability: Real-world constraints (timing, resources, budget) often require agile, imperfect, yet effective design strategies rather than rigid workflows.

Design Adaptability: Real-world constraints (timing, resources, budget) often require agile, imperfect, yet effective design strategies rather than rigid workflows.

Use Analogies to Inspire: When testing novel concepts, I found that providing relatable analogies (e.g., comparing our experimental concept to similar existing 3D design terms) and provided a range of options during sacrificial tests helped nudge clearer decisions. This approach revealed nuanced insights, particularly in areas like feature preferences and pricing strategies.

Use Analogies to Inspire: When testing novel concepts, I found that providing relatable analogies (e.g., comparing our experimental concept to similar existing 3D design terms) and provided a range of options during sacrificial tests helped nudge clearer decisions. This approach revealed nuanced insights, particularly in areas like feature preferences and pricing strategies.

Iterative Mindset: Working with engineers on a complex, industry-wide challenge reinforced the importance of iterative collaboration to deliver an MVP with core features, while saving "nice-to-have" features for future iterations.

Iterative Mindset: Working with engineers on a complex, industry-wide challenge reinforced the importance of iterative collaboration to deliver an MVP with core features, while saving "nice-to-have" features for future iterations.

Thanks for reading!

Every project tells my different quirks—would you like to see what’s next?

Danone SaaS