In a recent discussion, Brad Arkin, the Chief Trust Officer at Salesforce, highlighted the transformative potential of AI agents in security operations, distinguishing them from older automation technologies like robotic process automation. He emphasized that the implementation of AI agents marks the advent of a completely new technological landscape. This transformation is being approached with cautious optimism, starting with small-scale deployments to identify suitable use cases and gradually expanding to broader applications as insights are gained.
Presently, AI agents are playing a pivotal role at Salesforce by streamlining resource-intensive security tasks. These agents handle extensive documentation and routine processes, thus enabling teams to integrate these tools seamlessly into their workflows. Arkin acknowledged that while the AI might select incorrect data due to documentation discrepancies, such errors are attributed to the documents themselves rather than the AI agents.
The integration of AI agents into existing systems necessitates a cultural shift that Arkin believes should be approached with strategic planning. Rather than enforcing widespread operational changes, he advocates for a gradual incorporation of AI, where agents are initially used for drafting, and human experts refine the results. By opting for this phased approach, Salesforce aims to avoid potential cultural resistance that could hinder the deployment process.
In an interview at the RSAC Conference 2025, Arkin elaborated on the capabilities of AI agents to identify gaps and inconsistencies within workflows and documentation. He also discussed the importance of employing an internal reasoning engine to track and clarify the decisions made by AI agents. Furthermore, Arkin addressed the evolving role of security leaders in an infrastructure increasingly powered by AI, emphasizing the need for adaptation in leadership strategies.
Managing Salesforce’s security strategy, Arkin is dedicated to building trust with stakeholders and safeguarding customer data in an AI-centric environment. His extensive experience includes roles as the Chief Security and Trust Officer at Cisco and a 12-year tenure as the security chief at Adobe, equipping him with the expertise to navigate the challenges of integrating AI into security operations.