ASAPP’s GenerativeAgent: Transforming Airline Customer Service with AWS

ASAPP

By: Dhanya Kurup, Partner Solutions Architect – AWS
By: Nirmal Mukhi, VP & Head of Engineering, GenerativeAgent – ASAPP
By: Rachel Corbat, Partner Solutions Architect – AWS

The airline industry stands at a crossroads in customer service evolution. Airlines are seeing increased customer service demands as global air travel returns to pre-pandemic levels. Customers want immediate, personalized service across all channels, including mobile applications and social media platforms, airport self-service kiosks, and traditional channels like call centers and service counters. In turn, airlines need to manage operational costs and variable service demands. This transformation isn’t just about incremental improvements – it’s about reimagining customer experience through the lens of artificial intelligence and cloud technology.

ASAPP, an AWS Advanced Partner, provides AI-powered solutions for customer service. Founded in 2014, ASAPP helps transform contact centers by automating tasks, assisting representatives, and improving customer experience. Their GenerativeAgent platform handles complex conversations while maintaining enterprise-level performance and safety standards. This blog explains how ASAPP’s GenerativeAgent platform, built on AWS infrastructure, combines cloud capabilities with specialized AI to transform customer service operations. The solution uses Amazon Bedrock with Claude Sonnet and Claude Haiku foundation models, integrating with services like Amazon Connect for comprehensive capabilities. These include intelligent conversation management, human-AI collaboration, and enterprise integration.

The Evolution of Airline Customer Service

Research shows that existing phone-based customer service centers that rely on manual call routing and scripted responses are struggling to meet modern demands. These systems typically include basic Interactive Voice Response (IVR) menus, queue-based call distribution, and isolated customer relationship management (CRM) databases that don’t integrate with reservation systems. Customers expect faster and more convenient service while interactions grow increasingly complex.

Airlines must navigate sudden contact surges from weather events, time-sensitive customer needs, and international regulatory requirements. The complexity increases with the integration of multiple legacy aviation systems and processes. These include Passenger Service System (PSS) for reservations, Global Distribution Systems (GDS) for inventory, Departure Control Systems (DCS) for boarding, and Revenue Management Systems (RMS) for calculations. This fragmentation means customer service representatives often need to manually access and navigate multiple platforms to resolve even a single customer request.

Current self-service solutions have limitations. First-generation IVR systems and rule-based chatbots operate on pre-defined decision trees and fixed response templates. For example, when a customer calls about a delayed flight, these systems can only provide basic status updates but cannot handle complex requests such as multi-carrier rebooking or refund processing. Industry reports confirm these limitations lead to low containment rates, resulting in increased operational costs, decreased customer satisfaction, and added complexity to interactions, ultimately frustrating both customers and service representatives.

Transforming Self-Service Through AI

As per this Forbes report, customer experience improvements deliver measurable business value, driving revenue 4-8% higher than industry competitors. Companies with superior customer experience reduce service costs by 15-20% and generate 5 times more revenue than competitors.

ASAPP’s GenerativeAgent platform helps enterprises capture these opportunities through seamless integration with AWS services, maintaining existing technology investments while adding powerful generative AI capabilities.

Figure 1 – Workflow diagram showing ASAPP’s Multi Agent Architecture

Key Platform Capabilities

The GenerativeAgent platform transforms customer service with intelligent conversation management, human-AI collaboration, and enterprise integration. This diagram illustrates the step-by-step process of how customer interactions flow through the platform using AWS services:

Customer Interaction Initiation

• • Customer interaction begins when contacts arrive via voice using Amazon Connect or through digital channels
  • The system enables conversational AI-powered end-to-end interactions with seamless topic identification and switching through advanced natural language processing and contextual memory management
  • The platform maintains interaction history, manages backend API access, and leverages document context to enable complex problem-solving capabilities

Model Orchestration via Amazon Bedrock

• • The platform uses different foundational models for different tasks: Claude Haiku for input safety and scope validation, Claude Sonnet for reasoning, tool calling, and output verification
  • The precise choice of which models are used in the overall orchestration is influenced by the use case and customer preferences
  • Amazon Bedrock manages model deployment, model output, and scaling

System Integration Layer

• • Deep integration with existing airline systems through an adaptation layer that transforms legacy API formats and data structures into standardized formats that can be efficiently processed by AI models
  • Real-time access via APIs to operational data such as booking, flight status, and baggage enables GenerativeAgent to reason and solve customer problems
  • Secure retrieval and handling of customer information: API transforms redact sensitive data, and any exposed sensitive data is redacted prior to storage and encrypted in transit and at rest

Response Generation and Routing

• • The system intelligently routes tasks between AI and customer service representatives
  • Most requests including complex workflows can be handled with full automation
  • Human-In-the-Loop Agent (HILA™) functionality for complex scenarios

Continuous Learning and Optimization

• • Interaction outcomes feed back into the system
  • Continuous monitoring of system performance such as which conversations may identify opportunities for improvement, summaries, structured data (including sentiment analysis, resolution rates, and use case-specific metrics)
  • Automated scenario testing for chat and voice to simulate and evaluate system behaviors prior to production launch or to test changes

Leveraging Amazon Bedrock for Flexibility and Performance

The GenerativeAgent platform uses Amazon Bedrock foundation models with a built-in orchestrator that coordinates task execution. This flexibility optimizes responses by enabling specialized model selection: one model may handle empathetic dialogue, while another model focuses on transactional tasks like flight changes. The orchestrator coordinates tasks including input safety validation, reasoning, RAG processing, and output verification, while models used for each can be selected based on specific use case requirements. This approach ensures contact centers receive tailored AI solutions without performance trade-offs or additional security complexity.

Enterprise-Grade Security and Compliance

ASAPP’s platform maintains strong security practices and compliance frameworks including SOC 2, GDPR, and HIPAA requirements. The platform implements industry-standard encryption, data protection, and comprehensive access controls with audit trails. ASAPP provides privacy controls such as data redaction, access governance, monitoring, and incident response. This approach allows airlines to maintain robust data protection and demonstrate compliance while focusing on delivering exceptional customer experiences.

Real-World Impact: A Leading Airline’s Transformation

A major U.S. airline’s implementation of the GenerativeAgent platform demonstrates the solution’s transformative potential. Completed in 39 days, the implementation delivered impressive results:

These results showcase how AI-driven customer service can deliver tangible benefits while improving experiences for both customers and representatives.

Looking Ahead

ASAPP’s GenerativeAgent platform will continue to evolve with innovative AI/ML technologies and foundation models. Future enhancements will enable more predictive capabilities, cross-channel integration, and personalized customer interactions. Through this evolution, airlines can expect AI to anticipate and address customer needs while maintaining the balance between automation and human expertise.

Conclusion

ASAPP’s GenerativeAgent platform, built on Amazon Bedrock and AWS infrastructure, sets a new standard in customer service automation. For airlines and other industries, it offers enhanced customer experiences through AI that works with customer service representatives while maintaining enterprise security and scalability. This work between ASAPP and AWS shows how technology can transform challenges into opportunities for improved customer engagement and operational efficiency.

To learn more about ASAPP’s GenerativeAgent platform or to get started, contact ASAPP through the AWS Partner Network or check out their listing on AWS Marketplace.

.
.

ASAPP – AWS Partner Spotlight

ASAPP is an AWS Advanced Technology Partner and AWS Generative AI Competency Partner that provides AI-powered contact center software designed to transform customer service operations through advanced artificial intelligence and automation technologies that help organizations expand contact center capacity while reducing costs and improving customer experiences.

Contact ASAPP | Partner Overview | AWS Marketplace