Array of Mesages
Chat Dialog of Messages
Chat Dialog of Messages with Toggle Button for Human Interaction
Chat Dialog of Messages with AI Response
AI Response
Whatsapp Response
Email Sending
Architectural Diagram

Inspiration

Business communication is often fragmented. A customer may receive an SMS notification, contact support through WhatsApp, and receive a confirmation by email, with no shared context between those interactions.

We wanted to build a different experience where the conversation itself becomes the interface.

Messaging Experience was inspired by the idea of combining agentic AI with omnichannel communication, allowing customers to communicate naturally while an AI agent understands their intent, maintains context, and helps complete tasks across WhatsApp, SMS, RCS, and email.

What it does

Messaging Experience is an agentic AI communication platform that helps businesses have intelligent, context-aware conversations with their customers.

Customers can communicate through channels such as:

WhatsApp
SMS
RCS
Email

Instead of following rigid chatbot menus, customers can simply describe what they need.

For example:

"I want to reschedule my appointment to Friday."

The AI agent can understand the customer's intent, retrieve relevant context, use available tools to perform the required action, and respond with the result.

The platform combines AWS Communication Developer Services with AI agents to turn messaging from a notification channel into an interactive customer experience.

How we built it

Messaging Experience is built using:

React for the frontend
Python + Django for the backend
PostgreSQL for persistent data and conversation history
AWS AI Agents / Amazon Bedrock for agentic intelligence
AWS End User Messaging for SMS and RCS
AWS End User Messaging Social for WhatsApp
Amazon SES for email

The architecture separates the AI, communication, application, and data layers.

Customer
   │
   ├── WhatsApp
   ├── SMS
   ├── RCS
   └── Email
          │
          ▼
AWS Communication Developer Services
          │
          ▼
      Django API
          │
     ┌────┴────┐
     ▼         ▼
PostgreSQL   AI Agent
     │       Amazon Bedrock
     │          │
     │      Agent Tools
     │          │
     └────┬─────┘
          ▼
    Business Actions
          │
          ▼
AWS Communication Developer Services
          │
          ▼
       Customer

When a customer sends a message, Django receives the communication event, retrieves the relevant conversation context from PostgreSQL, and passes the necessary information to the AI agent.

The agent determines the customer's intent and can use controlled tools to perform business actions. Django validates and executes those actions before returning the result to the agent.

Challenges we ran into

Multi-channel communication

WhatsApp, SMS, RCS, and email all have different capabilities and message formats. Designing a common communication layer while still supporting channel-specific functionality was one of our main challenges.

Conversation context

An AI agent needs more than the customer's latest message. We needed a reliable way to maintain conversation history and customer context so that interactions could continue naturally.

Safe AI actions

We wanted the AI agent to perform useful actions without giving the model unrestricted access to the application.

We solved this by exposing controlled backend tools. The agent requests an action, while Django handles validation, authorization, and execution.

AI Agent
   ↓
Tool Request
   ↓
Django Validation
   ↓
Business Logic
   ↓
Database / Service
   ↓
Result

Creating a natural experience

We also had to move away from traditional chatbot flows and design interactions where customers can communicate naturally instead of navigating complicated menus.

Accomplishments that we're proud of

We are proud of creating a working architecture that brings agentic AI and AWS Communication Developer Services together in a single customer experience.

Some of our key accomplishments include:

Building an omnichannel messaging architecture
Integrating AI agents with communication workflows
Maintaining persistent conversation context
Designing controlled tools for AI-driven actions
Building a Django backend and React business dashboard
Creating an architecture that can support different industries and customer workflows

Most importantly, we are proud of shifting the idea of messaging from simply sending notifications to helping customers accomplish tasks through conversation.

What we learned

We learned that building an agentic communication platform requires more than connecting an LLM to a messaging API.

The AI provides intelligence, but the surrounding system needs reliable communication, persistent context, business logic, security, and controlled tools.

We learned to separate responsibilities:

React       → Business Interface
Django      → Application & Orchestration
PostgreSQL  → Conversation Memory
AI Agent    → Reasoning & Decision Making
AWS CDS     → Customer Communication

We also learned that giving AI agents access to well-defined tools is more reliable than allowing them to directly interact with application data.

What's next for messaging-experience

Our next goal is to make Messaging Experience more autonomous and capable of handling complete customer workflows.

We plan to add:

Intelligent channel selection
Proactive customer engagement
Personalised customer journeys
Multi-agent collaboration
More business tools and integrations
Conversation analytics
Voice and messaging experiences
Industry-specific AI agents
AWS Marketplace deployment

Our long-term vision is to make Messaging Experience an agentic communication layer for businesses, where customers can start a conversation in any supported channel and complete meaningful tasks without leaving the conversation.

The goal is simple: turn messaging from a notification channel into an intelligent customer experience.