MUSCAT: BankDhofar, the Second largest bank by branch network size in Oman, has engaged Reksoft, a multi-service technology company, to help develop a strategy for transitioning to a new IT architecture using an AI-first approach. This initiative is the first of its kind in Oman’s digital banking sector.

The application of AI significantly accelerated every stage of the project — from analyzing the existing architecture (spanning more than 70 systems and 1,000 integration flows) to creating a prototype of the target architecture. These results have laid the foundation for BankDhofar’s long-term competitiveness and digital agility.

The AI-first approach integrates artificial intelligence as a core element at every stage — from ideation to testing — ensuring that AI tools underpin processes rather than being applied sporadically. This method automates decision-making data generation, optimizes resource use, and accelerates development through analytics and machine learning.

As part of its wider digital transformation strategy, BankDhofar is transitioning from a monolithic architecture to a modern microservices platform. Reksoft, acting as technology partner, conducted a comprehensive review of the Bank’s current IT infrastructure, including system documentation, network configurations, and integration mechanisms. It also designed the target architecture and migration strategy for more than 70 services and applications.

The work was executed in two main stages. The team had to virtually reconstruct the current IT landscape from scratch, preserving both functional and non-functional requirements. In the first stage, Reksoft used large language models (LLMs) to reverse-engineer existing integration services and architecture, quickly restoring system logic. Retrieval-Augmented Generation (RAG) technology enabled the analysis of unsupported technologies and automated the creation of technical documentation. This produced a complete registry of systems and integrations, with details on data flows, protocols, and security levels.

In the second stage, specialists designed the target integration architecture using Apache Camel. AI assistants helped automate Python code generation for integration components, data processing routes, and architectural diagrams. Locally deployed LLMs evaluated the compatibility of existing systems with the new architecture, assessing technical complexity and business relevance. From this analysis, a migration roadmap was created, factoring in priorities and potential risks. A working prototype of the target architecture was also implemented, validating the chosen technology stack.

Commenting on the project, Dr. Tariq Saleh Mohamed Taha, Chief Information Officer at BankDhofar, said: “This project marks a strategic step in BankDhofar’s digital transformation journey, laying the foundation for a modern IT ecosystem that fosters innovation and accelerates time-to-market. While microservices architecture is now the global standard for digital business, Reksoft’s AI-first approach proved a true accelerator, delivering clarity and a concrete implementation plan.”

Alexey Lebedev, Director of the Banking and Finance Department at Reksoft, added: “Our deep industry expertise, combined with advanced AI tools, enables us to significantly improve the efficiency and transparency of complex IT processes. The AI-first approach can deliver more than 40% in resource savings — AI can achieve in hours what might take humans months, provided experts guide it with the right context. We thank BankDhofar for their trust and the opportunity to jointly set new standards for digital banking in Oman.”