AutoHome AI — Designing an Emotionally Intelligent Smart-Home Assistant

4 hours

AutoHome AI is a next-generation smart-home companion that blends intelligence, emotional awareness, and human-centered automation.

Overview

AutoHome AI helps individuals manage everyday household routines - from climate control to meal planning - through adaptive insights that learn from behavior, context, and mood.

Originally conceived as part of my Stanford UI/UX Design for AI Products capstone, AutoHome evolved into a high-fidelity prototype exploring trust-centered AI, explainability, and multi-device coherence across a modern smart-home ecosystem.

Problem

Smart assistants today often overwhelm users with control panels, fragmented apps, and non-transparent automation.
Interviews and surveys revealed three key frustrations:

  • Decision fatigue from managing repetitive household tasks.

  • Lack of contextual intelligence - systems don’t understand energy, emotion, or environment together.

  • Limited trust due to opaque automation and inconsistent feedback.

AutoHome AI set out to design an assistant that feels aware, calm, and collaborative - not controlling.

Research & Insights

conducted 3 interviews and 6 surveys with solo homeowners, neurodiverse users, and working parents.
Key insights shaped the design principles:

  1. Cognitive load is invisible : users value emotional tone as much as efficiency.

  2. Manual override builds trust : users want clear, reversible automation.

  3. AI must show its reasoning : transparency creates emotional comfort.

These insights informed the interface philosophy: ambient clarity, empathy in microcopy, and progressive disclosure of system decisions.


Solution

AutoHome AI organizes home management into one adaptive interface:

  • A dashboard that unifies all devices : Lights, Climate, Kitchen, Security, Music, EV, and Garden.

  • Routine Planner that syncs with mood and energy data to suggest or adjust schedules.

  • Sustainability Center to visualize energy and water usage with eco-suggestions.

  • Voice & Chat Mode that explains its reasoning and accepts feedback naturally.

  • Trust Dashboard allowing users to see and control what data powers automation.

Each feature aligns with Stanford’s Human-in-the-Loop AI framework - keeping the user central in decision-making.