Kodebaze Replaces 28-Year-Old Legacy Fleet Platform for Norwegian Vehicle Leasing Company in Nine Month
A 1998 Microsoft Access database with VBA forms — still processing live financial transactions when Kodebaze began the migration. 750,000 lines. 175 tables. Nine months to replace it safely.
AI modernization factory completes full migration from 1998 Access VBA database to modern Azure microservice architecture with zero production downtime
Wholesale replacement of a system this complex would have failed. We moved it piece by piece, one module at a time. That is the only approach that works on a legacy estate this old.”
COPENHAGEN, DENMARK, April 30, 2026 /EINPresswire.com/ -- Kodebaze, the AI modernization factory for large legacy software estates, today announced the successful completion of a nine-month modernization engagement for a Norwegian vehicle leasing company owned by a major European financial group. The project replaced a Microsoft Access database and VBA forms system built in 1998, running 750,000 lines of code across 175 database tables, with a fully modern microservice architecture on Microsoft Azure.— Claus Villumsen, CTO and Founder, Kodebaze
The legacy system had been processing live financial transactions continuously since 1998. When the client's parent company proposed a wholesale lift-and-shift cloud migration, Kodebaze recommended an alternative approach: incremental module extraction with continuous data synchronization back to the legacy system, keeping the business fully operational throughout the transformation.
"The standard advice is to plan exhaustively and then move," said Claus Villumsen, CTO and founder of Kodebaze. "Our experience is different. You need to understand enough to choose the right method, then move carefully, one piece at a time. Wholesale replacement of a system this complex would have failed. Incremental extraction worked."
The engagement also required a novel approach to AI-assisted code analysis. Kodebaze engineers first washed the 750,000-line VBA codebase — stripping framework artifacts, dead code, and library noise — before dividing what remained into semantic categories, each processed by a dedicated AI agent. This addressed a known limitation of large language models on legacy codebases: attention degrades across long sequences, making critical business logic buried deep in large files effectively invisible to the model.
One hard-won lesson emerged from the engagement. Changing business logic at the same time as refactoring the code structure is a mistake. When both change simultaneously, teams lose the ability to compare old behavior to new, and bugs become impossible to trace. Kodebaze now treats behavior preservation and structural refactoring as strictly separate phases on all engagements.
The project delivered a complete replacement of the legacy estate across all core modules: customer management, contract and quoting workflows, vehicle fleet inventory, invoicing, and integrations for vehicle data, insurance, and contract signing. The new platform runs on React, .NET, PostgreSQL, and Azure with full CI/CD pipelines and automated monitoring.
Key outcomes included complete elimination of a 28-year-old VBA and Access dependency, replacement of bare-metal on-premises infrastructure with managed cloud hosting, and zero downtime during the entire cutover period.
A detailed first-person account of the project, written by Villumsen, is available at kodebaze.com.
Kodebaze is the AI modernization factory for large, complex legacy software estates. Every refactor is governed, auditable, and reversible. Kodebaze delivers full codebase analysis in days, not months, with every change human-reviewed before it ships. Learn more at kodebaze.com.
Claus Villumsen
Kodebaze.com
+45 32 73 61 37
claus.villumsen@kodebaze.com
Visit us on social media:
LinkedIn
X
Legal Disclaimer:
EIN Presswire provides this news content "as is" without warranty of any kind. We do not accept any responsibility or liability for the accuracy, content, images, videos, licenses, completeness, legality, or reliability of the information contained in this article. If you have any complaints or copyright issues related to this article, kindly contact the author above.
