Architecting Trust: A Governance and Verification Framework for Autonomous AI Agent

TOPIC

AI agents are rapidly becoming part of the enterprise workforce, accessing systems, data, and business processes with increasing autonomy. Unlike traditional software, they can reason, make decisions, and act across multiple systems, creating new risks including privilege escalation, prompt injection, agent drift, and compliance failures.

This session introduces a practical Trust Framework for Autonomous AI Agents that enables organizations to innovate safely. Learn how to establish trusted agent identities, enforce least-privilege access, implement real-time guardrails, and maintain auditable accountability for AI-driven actions.

Through real-world scenarios and proven control patterns, assurance, cyber, and GRC leaders will gain actionable strategies to govern agentic AI using zero-trust principles, continuous monitoring, and human oversight, without slowing adoption.

As AI agents become your newest workforce, how do you ensure they remain secure, compliant, and accountable?

Learning Objectives

By the end of this session, attendees will be able to:

1. Identify key risks associated with autonomous AI agents, including prompt injection, privilege escalation, agent drift, and shadow AI.
2. Apply a four-pillar trust framework built on identity, authority, runtime controls, and auditability.
3. Implement guardrails and monitoring to detect and prevent unauthorized or unsafe agent actions.
4. Strengthen assurance and compliance through auditable AI operations, provenance tracking, and human oversight.

SPEAKER

Premkumar Thomas Augustine

Finvi, VP Engineering Services and CISO

Premkumar Thomas Augustine (Prem) leads Finvi’s Information Security, Corporate IT, Cloud Infrastructure, and Quality Engineering divisions as VP of Engineering Services and CISO. With over 25 years of IT leadership across FinTech, Healthcare, Telecom, and Consumer Tax, Prem specializes in embedding robust security governance into modern, high-complexity software environments. A recognized thought leader in the space, he recently spoke at RSA Conference 2026 on the emerging challenges of non-deterministic systems.

As CISO, Prem defines and executes Finvi’s enterprise security strategy, driving resilient protection across cloud infrastructure, product ecosystems, and workforce technology. His recent focus centers on establishing pragmatic trust frameworks, zero-trust architectures, and dynamic governance models that secure autonomous AI agents without stifling innovation. Prem regularly advises executive leadership on emerging security paradigms, risk vectors, and balancing operational freedom with uncompromising safety controls.

EVENT DETAILS

RATE DETAILS

Time: 5:00 PM - 6:15 PM PST

Location: Webinar

CPE: 1 CPE

Member: $0

Full-time Faculty/Student: $0

Non-Member: $10

CPE CREDIT & DISCLAIMER

ISACA reserves the right to alter or delete items from the program in the event of unforeseen circumstances. Material has been prepared for the professional development of ISACA members and others in the IT audit, control, security, and governance community. Neither the presenters nor ISACA can warrant that the use of material presented will be adequate to discharge the legal or professional liability of the members in the conduct of their practices. All materials used in the preparation and delivery of presentations on behalf of ISACA are original materials created by the speakers, or otherwise are materials which the speakers have all rights and authority to use and/or reproduce in connection with such presentations. All applicable copyrights, trade secrets, and other intellectual property rights in the materials are and remain with the speakers.

Please note: unauthorized recording, in any form, of presentations and workshops is prohibited.

Countdown