Natural-Language AI Layer for GDS Bookings, Refunds and Reissues

The project involved the development of a natural-language AI platform for simplifying booking and post-booking operations across global travel-distribution systems.

The solution was designed and developed by a SilwaTech partner company through SilwaTech’s partner delivery ecosystem.

The platform converted pasted emails, free-text instructions, and itinerary information into structured actions for

• Travel agencies and online travel agencies
• Tour operators and travel consortia
• Corporate travel departments
• Destination management companies
• Travel management companies

The solution supported bookings, refunds, ticket reissues, queue monitoring and operational reporting across web, mobile and messaging channels.

It was designed to make GDS operations more accessible while retaining validation controls before transactional actions were submitted.

Natural-Language GDS and PNR Builder

A conversational interface was developed to translate plain-language instructions into structured GDS operations.

Key capabilities included

• Interpretation of plain-English travel requests
• Extraction of booking instructions from emails
• Translation of natural language into structured GDS actions
• Creation and modification of passenger name records
• Cancellation of supported booking segments
• Management of flight, hotel, car and cruise segments
• Support for multi-segment itineraries
• Fare-rule validation before execution
• Application of relevant contract-rate information

Refund and Reissue Automation

Post-booking workflows were automated to reduce the manual effort required for refunds and ticket changes.

The platform provided

• Automated refund-eligibility checks
• Retrieval and interpretation of applicable fare rules
• Ticket-reissue preparation
• Automatic fare recalculation
• Enforcement of policy-based refund conditions
• Support for corporate travel policies
• Handling of NDC fares alongside traditional GDS content
• Structured exception handling for cases requiring specialist review

Queue Automation and Operational Insights

Continuous queue monitoring was introduced to reduce missed actions and booking losses.

Capabilities included

• Overnight monitoring of GDS queues
• Identification of pending and time-sensitive actions
• Alerts for items requiring agent attention
• Reduced risk of missed queue tickets
• Operational reporting through a central dashboard
• Finance-focused reporting for corporate travel teams
• Visibility of booking and servicing activity

Multichannel Access

The same natural-language workflows were made available across multiple user channels.

Access options included

• Web-based booking and servicing
• Mobile access to supported workflows
• WhatsApp-based booking requests
• Refund and reissue requests through messaging
• Integration through the WhatsApp Business API
• Consistent workflow handling across connected channels

The platform handled traveller records, booking information and transaction-related instructions, making security and operational control essential.

Key areas of focus included

• Role-based access for agents, supervisors and finance users
• Secure authentication across supported channels
• Fare-rule validation before transactional execution
• Controlled access to booking and traveller information
• Secure integrations with GDS and travel-content providers
• Traceable processing of booking and servicing requests
• Exception handling for requests requiring human review
• Data-handling practices designed to support GDPR requirements
• An AI-processing architecture designed to minimise persistent storage of sensitive information

The zero-storage design objective applied to information processed by the AI layer and remained subject to the retention requirements of connected booking systems, transaction records and applicable regulations.