Niloom

Project Overview

Client

Niloom AI

Industry

Generative AI · Spatial Computing · Creator Tools

Niloom AI is a generative AI platform that allows users to create AR/VR experiences using text or speech prompts. The web app serves as the core creation tool, enabling ideation, collaboration, editing, prototyping, and instant publishing to the Niloom Play app across devices like Meta Quest, iPhone, Vision Pro, and Android.

My Role

As co-lead UX designer for Niloom.ai's browser-based web app, I collaborated closely with a cross-functional team, including engineers, AI specialists, creatives, and product managers, to make advanced spatial computing accessible to both casual creators and professionals. This case study details the end-to-end process, from problem discovery to post-launch impact.

Core Challenge

How do you design a GenAI platform that is both powerful and playful. without intimidating creators?

AR/VR content creation traditionally demands specialized skills, expensive tools (e.g., Unity), and hardware expertise. excluding most creators. Workflows are fragmented, iteration slow, and cross-platform testing cumbersome.

Core pain points identified

Heavy developer dependency for prototypes
Disjointed asset sourcing (characters, audio, effects)
Complex publishing across VR/AR devices

Process/Methodology

I initiated the design process through collaborative ideation workshops, working closely with the CEO, engineers, AI specialists, and the creative team to define the UX vision of our generative AI platform. Together, I mapped the end-to-end journey of a system that transitions simple text prompts into fully immersive VR/AR scenes. customizable either manually or through our AI agent, NilooAI.

I moved quickly into low-fidelity explorations, producing rapid sketches to experiment with prompt flows, asset libraries, and real-time preview integrations. These early concepts helped us clarify how users would move from imagination to spatial output with minimal friction.

From there, I developed mid-fidelity clickable prototypes to validate the core journey: prompt → generate → edit → publish. These prototypes allowed us to test interaction logic, refine the editing experience, and ensure that AI-powered generation felt intuitive rather than technical.

The process was highly iterative, with frequent cross-functional reviews and early AI feature integration. particularly around handling ambiguous prompts through contextual suggestions, smart examples, and adaptive guidance. Throughout, I placed strong emphasis on responsive browser design and intuitive spatial metaphors, ensuring the experience felt both powerful and accessible across devices.

FigJam - Ideation Process