Transforming a Contact Center with Amazon Connect Customer

A Phased Deployment of Amazon Connect Customer Contact Center

D3Clarity’s approach to solving Accent on Languages’ challenges involved a carefully structured implementation plan, executed in three distinct phases. Each phase focused on progressively enhancing the contact center’s functionality while meeting the federal agency’s requirements. By leveraging Amazon Connect Customer—a cloud-based contact center service—D3Clarity was able to build a secure, flexible, and scalable solution tailored to Accent on Languages’ needs.

Phase 1: Launching the Core Contact Center

The first step was to deliver a Minimum Viable Contact Center (MVCC), enabling Accent on Languages to meet the federal agency’s baseline requirements. This phase laid the foundation for all subsequent enhancements.

Key Features Delivered:

1. Language-Specific Call Routing via DTMF Input:
   D3Clarity configured an Interactive Voice Response (IVR) system that allowed callers to select their preferred language using Dual-Tone Multi-Frequency (DTMF) inputs (i.e., pressing numbers on a keypad). Calls were routed to one of ten language-specific queues based on the input.
2. Agent Whisper Functionality:
   Once a call was assigned, the selected language was whispered to the agent via Amazon Polly’s text-to-speech capabilities, ensuring interpreters were immediately informed of the caller’s requirements.
3. Manual PIN Collection and Logging:
   To enable secure PIN collection, D3Clarity added a custom input field to Amazon Connect Customer’s Contact Control Panel (CCP). This allowed agents to manually enter the caller’s PIN during calls. Call metadata, including PINs, language selection, agent IDs, and call durations, were securely logged in Amazon DynamoDB.
4. Call Logging and Reporting:
   Basic reports were generated through Amazon CloudWatch, tracking key metrics such as call volumes, agent availability, and average wait times. Call logs were securely stored in DynamoDB, enabling easy access for operational reporting and billing.
5. Security and Compliance:
   To meet government compliance standards, all data—including PINs—was encrypted both in transit and at rest. Role-based access controls were implemented to restrict data access to authorized personnel only.

Outcome of Phase 1:

With these features in place, Accent on Languages successfully launched a fully functional contact center that met the federal agency’s immediate needs. The system was secure, scalable, and equipped with detailed logging for accurate billing and reporting.

Phase 2: Enhancing the Contact Center with Advanced Features

Building on the foundation of Phase 1, the second phase introduced advanced features to improve caller experience and increase operational efficiency.

Key Enhancements:

1. Voice-Based Language Selection with Amazon Lex:
   To enhance user experience, D3Clarity integrated Amazon Lex, an advanced conversational AI, enabling callers to select their preferred language using voice commands. This feature allowed for a more natural interaction compared to DTMF inputs. A fallback mechanism redirected callers to the original DTMF system in cases where voice input could not be processed.
2. Handling Unrecognized Languages:
   The D3Clarity team added functionality to handle unrecognized language inputs, guiding callers back to the original flow and ensuring continuity in call handling.
3. Gender-Based Agent Transfers:
   Although initially planned for Phase 3, D3Clarity began prototyping a feature allowing callers to request an interpreter of a specific gender, enhancing caller satisfaction and meeting certain cultural and personal preferences.

Outcome of Phase 2:

The enhanced contact center provided a smoother and more user-friendly experience. The voice-based language selection streamlined call routing, reducing caller frustration. This phase also demonstrated the potential for future flexibility, positioning Accent on Languages to adapt to additional client requirements.

Phase 3: Creating a Dedicated Development Instance and Knowledge Transfer

The final phase focused on long-term scalability and self-sufficiency for Accent on Languages.

Key Deliverables:

1. Dedicated Development Instance:
   The D3Clarity team set up a separate AWS Connect instance for development and testing. This allowed Accent on Languages to experiment with new features and configurations without risking disruptions to the production environment.
2. Gender-Based Transfer Options:
   A dual-queue system was developed to support gender-based routing. Agents could override gender preferences when necessary and update call attributes dynamically during calls.
3. Knowledge Transfer Sessions:
   D3Clarity conducted a series of knowledge transfer (KT) sessions to empower Accent on Languages’ team to manage their contact center independently. These sessions included:
   • Queue setup and management.
   • Instance replication for new clients.
   • Hands-on demonstrations, followed by team ownership of processes.
     All KT sessions were recorded for future reference, ensuring on Languages’ team could revisit the material as needed.
4. Minor Configuration Adjustments:
   As a final step, the team provided additional support to refine the system, including:
   • Adjusting agent permissions to improve visibility into call metrics.
   • Enhancing DynamoDB records to capture precise call start times.
   • Adjusting timeout settings to prevent session interruptions during longer calls.

Outcome of Phase 3:

With a dedicated development instance and comprehensive training, Accent on Languages gained the tools and knowledge to manage and expand their contact center independently. This phase ensured the long-term sustainability of the solution.