HomeNest CareBridge: Project Story

About the Project

HomeNest CareBridge is an intelligent customer care platform designed to bridge proactive rich messaging, conversational self-service, and cloud telephony into one continuous experience. Built on AWS Communication Developer Services and Amazon Connect, the system resolves post-delivery appliance issues through interactive RCS messages, dynamic DIY troubleshooting, photo-verified warranty parts dispatches, and context-preserving phone calls.

What Inspired Us

Post-purchase appliance support is broken and disjointed.
When an appliance fails, customers are forced to choose between rigid chatbots and long hold queues.
Every time a customer switches from chat to a phone call or a live agent, they face a painful context penalty.
Customers are forced to re-verify their identity and explain the issue all over again from scratch.
We built HomeNest CareBridge to eliminate this friction entirely.
The goal was to ensure moving from a text to a phone call preserves all customer and order details without repeating a single detail.

What It Does

Proactive RCS Engagement: Sends interactive post-delivery check-in messages with action buttons for Give feedback, I need help, and Call me.
Dynamic DIY Troubleshooting: Guides users through plain-language fault resolution, such as clearing washer E21 drain errors, directly inside the chat.
Instant Voice Bridge: Pressing Call me or dialing +1 281 670 7909 authenticates the caller via delivery ZIP code 78701 and resumes support immediately with full context.
Photo-Verified Spare Parts: Dispatches free warranty replacement hardware, such as wardrobe hinge packs, once defect photos are submitted via email.
Human-in-the-Loop Escalation: Enforces policy guardrails by automatically routing complex or restricted requests, like address changes and refunds, directly to human agents with complete transcripts.

How We Built It

Amazon Connect: Manages native phone numbers, conversational IVR contact flows, queue routing, and the unified agent workspace.
Amazon Lex and Amazon Bedrock: Handles natural language understanding and intent recognition across text and voice.
AWS Communication Developer Services: Delivers interactive RCS cards and automated email summaries through [email protected].
AWS Lambda and Amazon DynamoDB: Keeps track of session memory across chat and voice calls in real time.
Agent Workspace Integration: Embeds the HomeNest Care Desk inside Amazon Connect so agents see live call transcripts and customer history side-by-side.

Challenges We Faced

Real-Time Cross-Channel Sync: Ensuring the active troubleshooting step from an RCS message instantly transferred to the voice agent before the phone even started ringing.
Policy Boundaries Versus Automation: Allowing the AI to freely handle troubleshooting while setting strict rules that require human approval for financial transactions or address updates.
Secure Evidence Verification: Setting up an authenticated email workflow to cleanly validate photos of damaged parts before approving warranty replacements.

What We Learned

Context Continuity Builds Trust: Retaining customer details across transitions cuts resolution times dramatically and removes customer frustration.
All-in-One Architecture Wins: Running phone numbers, IVR, routing, and agent screens entirely within Amazon Connect avoids the headache of integrating multiple third-party tools.
AI Works Best With Humans in the Loop: The best support experience is not about replacing humans, but letting AI handle routine troubleshooting while giving human agents the full story when complex issues arise.