The session will cover: The agentic SDLC shift: How AI is moving beyond code assistance into autonomous planning, development, testing, and deployment—and why this changes the software trust model. Emerging risks and trust boundaries: Addressing excessive agency, prompt injection, poisoned context, insecure tool use, sensitive-data exposure, and software supply-chain risk. Secure-by-design agent architecture: Using hardened agent harnesses, sandboxed execution, least-privilege access, and governed tools and MCP integrations to contain risk. Accountability and verifiable assurance: Maintaining human oversight, policy-based authorization, end-to-end traceability, deterministic security checks, and runtime monitoring. Scaling securely without slowing innovation: Measuring trust and security outcomes while enabling organizations to adopt agentic development responsibly and at enterprise scale.

Attendees will leave with a practical trust model for agentic software development.
A reference set of preventive, detective, and recovery controls for coding agents.
Guidance for scaling agent adoption without creating unmanaged “shadow agent” ecosystems. A roadmap for moving from experimental AI coding tools to governed, observable, and enterprise-ready agentic engineering.