Agentic Modernization on Azure: A New Operating Model for Legacy Transformation
At Microsoft Build, the conversation around AI is no longer about isolated copilots or productivity gains and demos. It’s increasingly focused on real systems that can produce large scale transformation and accelerate complex enterprise modernization.
Mainframe modernization may be one of the clearest examples of where this shift matters.
For decades, modernizing core enterprise systems was treated as a long, sequential engineering effort: assess applications, document dependencies, rewrite code, migrate data, test extensively, then slowly cut over workloads.
Programs often stretched across years because the real challenge was never simply translating COBOL into another language. It was understanding and preserving the business logic embedded across millions of lines of interconnected systems.
Now, advances in agentic AI and cloud-scale orchestration are changing what is possible.
Together, Amdocs and Microsoft are applying agentic modernization patterns on Azure to help enterprises modernize mainframe estates faster, with stronger governance, greater automation, and continuous validation across the entire software lifecycle. The result is not just accelerated migration, but a fundamentally different operating model for enterprise transformation.
“The need for mainframe modernization is less about ability to function and more about capacity to evolve.”
The AI in mAInframe Modernization
Mainframes are still exceptionally resilient systems. Many continue to run mission-critical workloads in banking, telecommunications, insurance, government, aviation, and manufacturing.
But enterprise expectations changed. Organizations now need AI-ready data architectures, cloud-native integration models, real-time analytics, faster release cycles, an elastic infrastructure and modern developer experiences. Legacy environments were not designed for this level of adaptability.
At the same time, generative AI has exposed a deeper challenge: AI systems are only as useful as the accessibility and structure of the enterprise environment behind them.
In many organizations, the most valuable business logic remains trapped inside the legacy environment: COBOL applications, transaction systems, undocumented functionalities, proprietary data structures, or batch processing dependencies.
Traditional modernization approaches struggle because they rely heavily on manual discovery and fragmented tooling.
Let’s see how agentic AI changes that dynamic.
From Copilots to Coordinated Agents
One of the biggest shifts happening across software engineering is the transition from single-task assistants to orchestrated systems of agents.
In the joint Amdocs and Microsoft approach, modernization is no longer treated as a collection of disconnected tools. Instead, specialized agents collaborate across the lifecycle through a shared orchestration layer.
This includes agents for tasks such as mainframe discovery, dependency mapping, business capability analysis, documentation generation, quality validation, deployment orchestration and Day-2 cloud operations.
What matters is not simply that AI is executing these steps. The important shift is that these agents share context across workflows and continuously validate outcomes.
That changes modernization from a mostly sequential process into a continuously governed system.
“The entire process of mainframe application modernization can now be driven through agents.”
Azure as the Runtime for Agentic Modernization
What makes this model operationally viable at enterprise scale is Azure.
Modernization workflows require massive parallel execution, secure orchestration, scalable inference, governance, and integration across complex enterprise environments.
The joint Microsoft–Amdocs architecture combines Azure’s AI and cloud foundation with the Amdocs Agentic Services platform, which acts as the operational orchestration and execution layer for large-scale modernization programs.
While Azure provides the scalable AI, data, and cloud infrastructure, Amdocs orchestrates end-to-end modernization flows, domain-specific agentic workflows, governance, and operational execution across enterprise application estates.
The joint architecture combines:
- Azure AI Foundry for scalable AI execution and model orchestration
- Microsoft Fabric for enterprise data landscape analysis
- Azure-native CI/CD, provisioning, observability, and governance services
- Azure SQL for modernized transactional environments
- Amdocs Agentic Services as the system of action orchestrating modernization workflows across cloud, application, testing, operations, and data domains
- Amdocs domain-specific agents and orchestration frameworks enabling governed, explainable, and operationally accountable transformation flows
- Integrated governance, security, and operational controls spanning both Azure and Amdocs operational layers
This enables modernization programs to run across thousands of applications and dependencies without requiring teams to manually provision infrastructure or coordinate every transformation workflow individually.
Context Is the Real Modernization Challenge
One of the recurring misconceptions around AI modernization is the idea that source code alone is enough. It is not. Enterprise modernization depends on reconstructing operational context:
- Why systems behave the way they do
- Which integrations are business critical
- How transaction flows operate
- Which dependencies are implicit
- What regulatory constraints exist
- How data behaves in production
The joint Amdocs and Microsoft framework addresses this through layered context reconstruction.
LLMs are enriched using documentation, dependency graphs, semantic application models, runtime analysis, system inventories, Microsoft Fabric insights into enterprise data landscapes.
This allows agents to move beyond simple code translation and toward actual system understanding. The result is modernization decisions driven by business context, not just syntax conversion.
Modernization as a Foundation for AI-Native Enterprises
The long-term significance of mainframe modernization is not simply infrastructure reduction but also preparing enterprise systems for the next generation of AI-native operations. Once applications and data are modernized on Azure data becomes more accessible, AI models can operate on unified architectures, analytics become real-time and automation scales across workflows.
As a result, innovation cycles accelerate. Mainframe modernization becomes the foundation for continuous transformation, not the endpoint.
Learning how it can be done is ultimately what Build is about. If you want to go deep on real code, real systems, and real workflows with teams building and scaling AI, this two-day conference is for you.
The future of enterprise AI is not only about new applications but also about unlocking the systems enterprises already depend on. And increasingly, that requires agentic modernization operating at cloud scale.
Learn more about the Amdocs & Microsoft Agentic AI Modernization solution: download the Value-Driven Mainframe Modernization, the Agentic Way Whitepaper.
Automated mainframe modernization
Unlock agility, innovation and all the cloud benefits while eliminating mainframe-related risks with an end-to-end, AI-driven transformation solution.
Value-Driven Mainframe Modernization, the Agentic Way
Deliver faster, safer, AI-ready modernization with Microsoft Azure and Amdocs Agentic Services
