Spark
Matching potential brand partners and developing collaboration directions into proposals that can be discussed and revised.
- My role
Team project; architecture, frontend development, interaction and visual design, and COLLIDER integration
- Year
2026
- Context
Team collaboration at AIxORigin hackathon
- Status
Award-winning prototype
Project context
Spark is an application for exploring brand partnerships and creating proposals together. It organises brand resources, needs, audiences and collaboration intentions, presenting potential partners through a gravity graph and card-draw interface. Users can examine matching reasons, complementary conditions and information still to be confirmed to understand possible directions for cooperation.
After a partner is selected, both brands’ information and the collaboration brief enter the COLLIDER canvas. Conversation supports generating and revising copy, visuals and channel materials, connecting brand discovery with proposal production. Brand characters and a directory help distinguish brands, while the canvas retains discussions, materials and revisions.
The project was completed collaboratively by a team during the AIxORigin hackathon and received the Special Pioneer Innovation Team Award in the Agent Track. The current result is a working prototype with simulated rule-based matching, demonstrating the complete workflow from brand information and partner discovery to collaborative proposal development.
Scope and limitations
Local prototype. Matching uses mock rules, not commercially validated recommendations. AI understanding and generation require configured services; connection requests and mutual confirmations are local demo states and send no real invitations. Nailong × Cotti is a non-commercial workflow demonstration, not evidence of a real partnership, official authorization or a definitive match.