Download now

Executive Summary

This white paper focuses on mainframe application estates and shows how Gemini together with CAST software mapping & intelligence gives teams a structure-aware view of COBOL/PL/I code, JCL flows, datasets, and surrounding integrations. The result is faster, safer modernization - clear scoping, evidence-based choices, and LLM workflows grounded in the real architecture.

Learn about:

  • Scaling insight across large mainframe portfolios by turning source and JCL into navigable architecture maps and end-to-end dependency views.
  • An agentic approach that blends AI-generated documentation and vector retrieval with CAST structural analysis to answer code-to-architecture questions with evidence.
  • Practical outcomes for teams: precise impact analysis, tighter test scoping, defensible rehost/replatform/refactor decisions, and clear refactoring paths.

The white paper distills a joint Google - CAST proof of concept and a repeatable approach - from hierarchical documentation and retrieval to CAST’s structural analysis and multi-agent orchestration - so organizations can bring fact-based visibility to complex mainframe portfolios and accelerate modernization at enterprise scale.

Philippe GuerinCAST
Philippe Guerin

CTO North America

Philippe Guerin is CTO North America at CAST and leads the team of CAST Solutions Architects. He is an expert in application modernization, architectural design, and software intelligence. Over the last 25 years Philippe has been helping businesses, government agencies, management consultancies, cloud vendors and systems integrators assess large IT organizations application landscapes and accelerate transformation efforts. You can contact Philippe directly at p.guerin@castsoftware.com.

Cat PerryGoogle Cloud
Cat Perry

Business Growth Specialist

Cat Perry is a Business Growth Specialist for the Mainframe Modernization team at Google Cloud. She drives partner relationships with Independent Software Vendors and Global System Integrators focused on mainframe modernization. A key achievement is her leadership in the global launch of Google's Dual Run program, enabling customers to test and de-risk their modernization journey from the mainframe to Google Cloud. Over the past several years, she has been instrumental in growing the practice's brand and enablement of Google solutions for legacy transformation, working closely with technical colleagues to guide customer modernization success. You can contact Cat at catperry@google.com.

Intelligent Mainframe Modernization with Gemini & CAST

A joint perspective from CAST and Google Cloud on modernizing mainframe and midrange estates by combining CAST’s deep structural blueprinting with generative artificial intelligence guidance (Gemini) to move from analysis to actionable modernization strategy.

Explain why mainframe modernization is hard (interwoven legacy technologies, high maintenance cost, fewer experts, tight coupling with customer-facing systems).

Use CAST Imaging to trace end-to-end transactions and map precise program and data paths (for example, from web entry points through transaction managers into business logic and databases).

Use Gemini (fed with CAST Imaging context via an integration server) to recommend modernization paths and accelerate the shift from blueprint to transformation plan.

Provide fragmentation strategies grounded in CAST dependency and coupling analysis (data-centric clustering, low-coupling “silos” for pilots, exposing core functions as modern application programming interfaces).