Insights AWS case study

Production Enterprise AI Customer Service Platform

Bexprt was selected to design and deliver a production-ready Enterprise AI Customer Service Platform on AWS, transforming SAT’s customer-service experience. Built on Amazon Bedrock, the platform orchestrates four AI agents to provide 24/7 multilingual self-service in Arabic and English across bookings, refunds, luggage and trip enquiries, fully integrated with SAT’s CRM and operational workflows.

As part of the production architecture, Bexprt uses Anthropic Claude Sonnet and Claude Haiku via AWS within an evaluation pipeline for customer-service chats. Claude provides an additional AI layer for evaluating and assessing customer interactions within the overall customer-service solution, helping Bexprt maintain the quality and effectiveness of AI-generated conversations in production.

Measurable Business Impact

Delivered in just 12 weeks, the 100% serverless production solution now resolves more than 80% of customer enquiries autonomously and has contributed to an approximately 60% reduction in customer-service support costs. The deployment has enabled SAT to move from a predominantly manual customer-service model to a scalable, AI-enabled self-service operation available 24/7 across Arabic and English.

The SAT deployment represents the first phase of a broader AI transformation programme between SAT and Bexprt, with further opportunities being explored across areas including Voice AI, Contract Intelligence, Yard Inspection and other enterprise AI use cases.

The Challenge

SAT (SAPTCO ALSA for Transportation) is one of Saudi Arabia’s leading intercity transportation operators, serving customers across a wide range of journeys and service enquiries.

Following the modernization of its technology infrastructure on AWS, SAT wanted to go further and transform its customer-service operation. The objective was to reduce reliance on manual and outsourced support, provide customers with faster 24/7 assistance in both Arabic and English, and create a scalable foundation for further AI-driven innovation.

The challenge was not simply to introduce a chatbot, but to build a production-ready Enterprise AI platform capable of handling real customer journeys, integrating with SAT’s existing systems, and operating securely and reliably at scale.

The Bexprt Approach

Working closely with SAT and AWS, Bexprt first established a modern, secure, highly available and resilient AWS environment that could support SAT’s ongoing digital transformation.

Building on this foundation, Bexprt designed and delivered a production-ready Enterprise AI Customer Service Platform on AWS in just 12 weeks.

Built on Amazon Bedrock, the 100% serverless platform orchestrates four AI agents to provide 24/7 multilingual self-service in Arabic and English across customer journeys including bookings, refunds, luggage and trip enquiries, with integration into SAT’s CRM and operational workflows.

As part of the production architecture, Bexprt also uses Anthropic Claude Sonnet and Claude Haiku via AWS within an evaluation pipeline for customer-service chats. Claude provides an additional AI layer for evaluating and assessing customer interactions within the overall solution, helping maintain the quality and effectiveness of AI-generated conversations in production.

From Migration to AI Innovation

The real transformation comes in Phase 2, where SAT is leveraging Amazon Q Business and Amazon Q in QuickSight to revolutionize customer service and data analytics. This journey from infrastructure modernization to AI implementation demonstrates what’s possible when organizations take a strategic, phased approach to digital transformation.

From Migration to AI Innovation

SAT’s journey demonstrates how cloud modernization can provide the foundation for moving Enterprise AI beyond experimentation and into real production operations.

The AI Customer Service Platform is now deployed in production and autonomously resolves more than 80% of customer enquiries, significantly reducing the volume of interactions requiring third-party call-centre support.

As SAT stated during its customer presentation, deployment of the solution resulted in a 60% saving in third-party call-centre costs.

The production deployment combines Amazon Bedrock, AI agents, enterprise data, CRM integration and Anthropic Claude within the evaluation pipeline to create a scalable customer-service operation available 24/7 across Arabic and English.

This represents the first phase of a broader AI transformation journey between SAT and Bexprt, with further opportunities being explored across areas including Voice AI, Contract Intelligence, Yard Inspection and other enterprise AI use cases.

Looking Ahead

SAT’s transformation represents a broader trend we’re seeing across the Kingdom of Saudi Arabia – organizations are ready to move beyond proof-of-concepts to enterprise-scale AI deployments.

By combining AWS’s powerful AI services with Bexprt’s implementation expertise, we’re helping Saudi businesses unlock new levels of efficiency and customer experience.

Recap: Bexprt x SAT at AWS Cloud Day Riyadh 2025

On September 29th at the Hilton Riyadh, Bexprt Founder Mo Hamdy joined SAT’s CTO Bader Al-Balwiy and AWS’s MD | GM Mid-Market Enterprise EMEA Amélie Clugnet for a panel discussion on “AI chatbots and Amazon Q data visualization: SAT’s journey.” The session explored how organizations can move from AI experimentation to production-ready solutions that deliver measurable ROI.

Yousef Allahham, Bexprt Solutions Architect, also presented on The Evolution to Agentic AI, and how enterprises can bridge the gap between innovation and reality.

Discover more

Bexprt ◦ Your Enterprise AI & Resilience Partner. hello@bexprt.com