Aura: The AI Co-Pilot for Social Work

HCI & AI Integration Course • Individual Project Diary

Designing Augmented Intelligence for High-Stakes Human Services

1. Project Description & My Goal

Aura is an ambient AI co-pilot for social workers (e.g. Maya in Hong Kong SWD) conducting high-stakes home visits. It tackles post-pandemic challenges: surging caseloads, manual documentation overload, personal safety risks, and burnout (1.7× surge).

Key features include hands-free ambient narrative capture, visual family ecosystem mapping, explainable AI triage, real-time safety guardian, wellbeing monitoring, and smart planner — all with strict human-in-the-loop control.

My Goal:

"A dedicated social worker needs a way to remain fully present with clients while seamlessly capturing data and ensuring safety, because their current cognitive load leads to burnout."

User Profile Maya

2. My Responsibilities in the Group & Achievements at Each Milestone

Phase 1
Research & Needfinding – Led user interviews with social workers and delivered POV statement + needfinding artifacts.
Phase 2
Ideation & Prototyping – Drove brainstorming and built interactive prototypes for ambient capture and visual mapping.
Phase 3
Evaluation – Coordinated expert validation; achieved 90% reduction in reporting time and strong ethics validation.

3. My Learning and Execution Process

Needfinding → Solving Real Pain Points

Conducted contextual research and engaged directly with social workers to uncover safety risks, cognitive overload, and admin burden.

Needfinding

Quick Prototyping & Iteration

Rapid low-to-high fidelity prototyping in Figma + working video demo. Refined features after every round of feedback from social workers.

Evaluation & Refinement

Conducted speed-dating sessions and expert reviews → iterated on explainable AI and privacy features.

Evaluation

4. Personal Reflection

This project taught me the power of NEEDFINDINGS in the real world. By talking directly to social workers, I learned how to move beyond assumptions and truly solve painful problems like burnout and safety risks.

The **quick prototyping → feedback loop** was the biggest skill I gained. Seeing my initial ideas evolve dramatically after user testing showed me how iteration turns good designs into great ones.

Key lessons learned:

  • AI in human services must be ETHICS FIRST — human-in-the-loop and explainable AI are essential.
  • Ambient computing dramatically reduces cognitive load and lets practitioners stay present with clients.
  • Collaboration and rapid iteration are more important than perfect first drafts.

"Aura reminded me why I love HCI: technology should amplify human connection and compassion, never replace it."

— Nok Chan, May 2026