ParentBridge AI — One Conversation, Every Channel

• 

  IMAGE

Inspiration

School-family communication is fragmented across email and text messages. Parents repeat the same questions when switching channels, and staff lose context. ParentBridge AI explores an easier, consent-aware experience.

What it does

The project provides a responsive school communication dashboard that combines SMS and email into a shared conversation timeline. Its prototype supports family communication preferences, consent checks, opt-out commands, school resource questions, conversation continuity when switching channels, and human escalation for sensitive or consequential requests.

How we built it

Python powers the backend and its conversation orchestration, with browser-native HTML, CSS and JavaScript for the dashboard. The codebase includes AWS SAM infrastructure for API Gateway, Lambda, DynamoDB and SNS; Boto3 runtime adapters implement Amazon SES v2 send_email and AWS End User Messaging SMS send_text_message; Amazon Bedrock Converse is an optional agentic response path when configured. An offline demo uses deterministic answers and simulated delivery without AWS credentials.

Testing and limitations

The local prototype passed 19 automated tests, including mocked AWS SDK invocations, cross-channel continuity, consent enforcement and escalation. The source package contains infrastructure and deployment instructions. An actual AWS deployment and live CDS sends have not yet been completed or verified. The local demo and video illustrate simulated operation, not live AWS traffic. The system uses fictional family information and is not production-ready for real school records.

Challenges

Persisting a coherent thread across phone and email identifiers while enforcing consent, deduplicating events, and limiting automated responses to safe cases.

Accomplishments

A working local operator dashboard, documented AWS service adapters, infrastructure-as-code, an architecture diagram, and an approximately three-minute narrated demo of the offline prototype.

Next steps

Deploy the SAM stack to an authorized AWS account; verify SES/email and eligible SMS originators; test real CDS API calls with consented recipients; capture AWS runtime evidence; validate the entrant's APN and ACE opportunity details before submitting; the source repository and YouTube demonstration are already available.

Source, Architecture, and Demo

Public source code, MIT license, and reproducible instructions Architecture diagram (SVG) Deployment and live AWS verification instructions Three-minute video demonstration

Runtime disclosure: The linked video demonstrates an offline simulation. Code includes AWS service adapters, but actual AWS deployment and live sending with AWS CDS have not been verified. The architecture diagram is linked in the writeup because direct file attachment remains pending.

Project resources (public URLs): GitHub source and MIT license: https://github.com/samertall/main/tree/main/parentbridge-ai Architecture diagram: https://github.com/samertall/main/blob/main/parentbridge-ai/assets/architecture.svg AWS deployment instructions: https://github.com/samertall/main/blob/main/parentbridge-ai/docs/DEPLOYMENT.md Demo video: https://www.youtube.com/watch?v=DsT6T3kpu90

Submission disclosure: The local video is a simulation; actual deployment on AWS and live sending through CDS have NOT been verified. These external links do not replace any required Devpost file upload.

Built With

• amazon-bedrock
• amazon-dynamodb
• amazon-ses
• amazon-sns
• aws-end-user-messaging
• aws-lambda
• aws-sam
• boto3
• css
• html
• javascript
• python