JUN WANG

Case study

AI Cloud Eye Smart Community

This project focused on designing a workflow product for property and security teams, using AI recognition to support patrol automation, incident tracing, and exception handling while preserving operational fallback when the system could not resolve a case automatically.

Role
Product manager with workflow and interaction ownership
Users
Property staff, security teams, and management roles
Focus
AI-assisted operations, multi-role workflow design, dashboard and mini-program flows

The problem

Smart community products often promise automation, but operational teams still face messy real-world conditions. Vehicle or incident recognition may be incomplete, dangerous-area alerts may need review, and frontline staff need a way to move from detection to action quickly.

The challenge was to make AI useful inside a real operating workflow rather than treating recognition itself as the product.

What I worked on

  • Framed the product problem around operational efficiency, not just algorithm capability.
  • Mapped workflows across web management views and mobile or mini-program actions.
  • Defined dashboard structure to support monitoring, review, tracing, and follow-up.
  • Designed fallback interactions for cases where AI recognition was incomplete or uncertain.
  • Coordinated delivery so the product logic matched real operational behavior.

Key decisions

I treated the system as a multi-role workflow rather than a single-screen management tool. Property, security, and management users needed different levels of visibility and action.

I also made room for manual correction. A strong AI operations product cannot assume the machine is always right. Good fallback design keeps the workflow usable when automation is incomplete.

Finally, I focused the experience on decision speed. Alerts, traces, and follow-up actions needed to feel actionable, not merely informative.

Why this case matters

This case shows my ability to work on a commercial system where AI is only one layer of the solution. The harder part was structuring the surrounding workflow so teams could actually use automation in daily operations.

It demonstrates systems thinking, multi-role interaction design, and the practical judgment needed to balance AI capability with human operations.