Agentic AI
,
Artificial Intelligence & Machine Learning
,
Events
Salesforce’s Brad Arkin on How Agents Are Transforming Security Ops
Brad Arkin, chief trust officer at Salesforce, cautioned against equating AI agents with past automation trends such as robotic process automation. “All of this that we’re doing is completely new technology. We are starting out with small steps and looking for places where we think it’s going to make sense to deploy, and then trying out these use cases, learning from that, and growing in other directions,” he said.
See Also: Unlocking Enterprise Productivity and Innovation Through Secure Agentic AI
Currently, agents are assisting Salesforce in routine but resource-heavy security processes – from heavy documentation, repetitive processes to boilerplate material. Teams are being encouraged to absorb these tools naturally into their workflows. “If document A says its eight hours and document B says four hours, they disagree. The agent may pick the wrong one – but that’s a documentation error, not the agent’s fault.”
Culturally, the shift requires careful planning. Arkin recommended avoiding forced, large-scale operational changes. Instead, he suggested a phased approach in which agents generate first drafts and human experts refine and finalize the work. “We’re not picking any cultural battles upfront. That could slow out the deployments.”
In this video interview with Information Security Media Group at RSAC Conference 2025, Arkin also discussed:
- How AI agents uncover gaps and inconsistencies in workflows and documentation;
- The role of an internal reasoning engine to track and clarify agent decisions;
- The evolving responsibilities of security leaders in an AI-powered infrastructure.
At Salesforce, Arkin oversees the company’s security strategy, builds trust with stakeholders and protects customer data in an AI-first world. Previously, he was chief security and trust officer at Cisco and served as the security chief at Adobe for 12 years.