The client is a global automotive manufacturer with large scale vehicle production and sales networks across North America, Europe, and Asia.
IBM Consulting is IBM’s professional services organization, helping enterprises plan and deliver transformations from strategy and architecture through modernization, implementation, and operations, with around 160,000 consultants worldwide.
legacy documentation
impact analysis
gain with grounded agents
Ted Trichew
A global automotive manufacturer launched a multi year modernization of its legacy application applications so capabilities could be reused worldwide and integrated with a modern packaged CRM platform.
The scope included a very large legacy Java application of about 10 million lines of code, with complex dependencies and limited up to date documentation. Knowledge lived with a small set of experts, making impact analysis slow and changes risky for critical business flows.
IBM Consulting began with deterministic discovery for the Java estate. CAST Imaging was used to produce architecture maps, call graphs, and business tagging, then exposed these outputs via API into a graph database and vector store.
IBM layered a conversational interface and autonomous agents on top of this knowledge base to support demystification, impact analysis, refactoring decisions, and modernization planning at scale.
With CAST Imaging providing deterministic code intelligence, IBM’s AI agents were grounded to deliver consistent, complete answers at 90%+ accuracy instead of relying on tribal knowledge.
IBM reported up to 70% less time spent documenting legacy code and about 60% less time on impact analysis, roughly 3.3x and 2.5x faster.
Agentic workflows pushed productivity above 50%, accelerating delivery while reducing change risk and rework.