To address these requirements, OneData built the OneConnect platform using Amazon Connect and a range of AWS AI, communication, analytics, and infrastructure services. 

The platform operates across four Amazon Connect instances, creating a governed foundation for supporting multiple customers. Tenant-specific configurations, workflows, identity controls, and contact center resources help maintain separation between customer environments. 

SAML identity federation was implemented to provide secure enterprise access. AWS Cloud Development Kit was used to automate infrastructure provisioning and standardize the deployment of tenant environments. 

This automated approach reduces manual configuration effort and supports more consistent customer onboarding. 

Autonomous Voice AI with Amazon Nova Sonic 

Amazon Nova Sonic was integrated to enable autonomous and natural-language voice interactions. 

The voice AI connects through Amazon Kinesis Video Streams and custom AWS Lambda bridge functions. This allows the AI agent to receive live audio, understand customer requests, generate contextual responses, and continue the conversation in real time. 

The autonomous voice agent can handle inbound calls from beginning to end and escalate customers to human agents when additional support is required. 

This enables OneConnect customers to automate common voice interactions while preserving human involvement for complex or sensitive conversations. 

Tenant-Specific Self-Service with Amazon Lex 

Ten Amazon Lex V2 bots were implemented to support tenant-specific customer self-service. 

Each bot can be configured according to the workflows, terminology, requirements, and services of the relevant customer. This enables organizations to deliver differentiated conversational experiences while operating on the shared OneConnect platform. 

Amazon Lex supports intent recognition and automated customer interaction flows, helping resolve common inquiries before human agent involvement becomes necessary. 

Real-Time Agent Assistance with Amazon Q in Connect 

Amazon Q in Connect was implemented to provide real-time assistance during customer conversations. 

The solution uses three knowledge bases and an Amazon Aurora vector store to retrieve relevant information based on the context of each interaction. 

Amazon Q in Connect surfaces appropriate guidance and information to agents while they are communicating with customers. This reduces the need to search manually across multiple resources and helps agents provide more accurate and consistent responses. 

Conversational Analytics and AI-Generated Summaries 

Amazon Connect Contact Lens and a Live Call Analytics architecture were implemented to provide real-time transcription, sentiment analysis, and conversational insights. 

Customer interactions are transcribed and analyzed, providing greater visibility into conversation quality, customer sentiment, agent performance, and recurring service patterns. 

Amazon Bedrock generates conversation summaries, reducing the need for agents to manually document complete interactions and making important conversation details easier to review. 

Together, Contact Lens and Live Call Analytics provide conversational analytics across supported customer contacts. 

Omnichannel Customer Engagement 

OneConnect unifies multiple communication channels within a single contact center platform. 

The platform supports: 

• Voice 

• Apple Messages for Business 

• SMS 

• Web chat 

• WebRTC 

• LiveKit-powered communication 

These channels are presented to agents through a custom Contact Control Panel, Flutter mobile applications, and AWS Amplify web applications. 

This unified experience enables agents to manage voice and digital conversations from one operating environment while allowing customers to communicate through their preferred channels. 

Integrated Case Management 

Amazon Connect Cases was implemented to help agents track customer requirements, issues, actions, and resolution progress. 

Case records preserve context across different interactions and communication channels. When customers return or move between channels, agents can review previous conversations and continue the engagement without unnecessarily repeating questions. 

This provides a more consistent customer experience and gives agents a clearer view of the customer journey. 

Legacy 3CX Migration 

OneData migrated a legacy 3CX telephony deployment to Amazon Connect using a phased, parallel-run approach. 

The legacy environment and the Amazon Connect platform operated simultaneously during the migration. This allowed OneData to validate telephony functions, routing, contact flows, and agent operations before completing the transition. 

The 3CX configuration was exported and stored in Amazon S3 to provide a validated rollback option if restoration became necessary. 

This phased approach enabled the migration to be completed with zero per-tenant downtime while preventing the transition from affecting other customers operating on the OneConnect platform. 

By combining Amazon Connect, Amazon Nova Sonic, Amazon Lex, Amazon Q in Connect, Amazon Connect Cases, Amazon Connect Contact Lens, Amazon Kinesis Video Streams, Amazon Bedrock, AWS Lambda, Amazon Aurora, Amazon S3, AWS Amplify, and AWS Cloud Development Kit, OneData established a unified and scalable contact center architecture for OneConnect.