Agentic Software Development

Integrate AI agents into enterprise software development with governance and control Agentic Software Development introduces a systematic framework for deploying specialized AI agents in distinct development roles—architect, security engineer, code reviewer—that collaborate to produce version-controlled, secure, auditable applications. Written by Priyanka Vergadia, Senior Director of Developer Advisory at Microsoft, and Shree Chinnasamy, a Senior Customer Experience Engineer in Microsoft Azure, this book delivers the implementation guidance technology leaders need to shift from passive AI usage to active AI orchestration. The book provides a blueprint for assigning AI agents to specific software development functions while maintaining traceability, compliance, and architectural integrity. Readers learn to measure ROI on AI-driven development, manage technical debt introduced by AI coding tools, and build governance structures that keep codebases auditable rather than opaque. Each chapter addresses the practical tension between accelerating development cycles and preserving the security and control enterprises require. Readers will also find: A framework for transitioning AI from a productivity tool to a structured, role-based enterprise development partnerStrategies for maintaining architectural integrity and preventing enterprise codebases from becoming ungovernable black boxes over timeGuidance on building compliance and security controls directly into AI-driven software development workflows from the startTechniques for measuring return on investment and quantifying the broader business impact of agentic AI adoption initiativesMethods for leading engineering teams through the organizational shift from manual code review to AI-orchestrated development processesAgentic Software Development serves CTOs, software architects, engineering directors, and senior developers integrating AI agents into enterprise workflows. It also supports management consultants, technology advisors, and computer science educators who need to understand enterprise-grade AI development practices to advise clients, assess technical risk, or prepare students for modern software engineering roles.