AI agents are appearing across every enterprise platform, but most still struggle to move beyond scripted automation into systems that can reason, adapt, and operate within real workflows.

On this episode of Ctrl + Alt + AI, host Dimitri Sirota sits down with Vivienne Wei, COO for Unified Agentforce Platform, Apps & Industries Technology at Salesforce, to examine what it actually takes to deploy agentic AI at scale. Vivienne leads the unified Agentforce platform, which brings together data, governance, and AI to enable agents that can act autonomously across enterprise systems.

She explains how agentic AI differs from earlier automation, why context engineering is becoming a core requirement, and how governance models must evolve as agents become active participants in business processes. 

For security and data leaders, this discussion highlights a shift already underway. Agents are not just interacting with data. They are acting on it, which raises new questions around access control, accountability, and trust.

What to expect:

• Agentic AI requires governance models built for non-human actors

• Context engineering determines whether agents are useful or risky

• How to start with business outcomes, not agent capabilities

Things to listen for: 

(00:00) Meet Vivienne Wei

(01:25) What Agentforce is and how it works

(02:35) Defining agentic AI vs traditional automation

(04:24) Why context engineering is becoming critical

(06:00) Governing agents as non-human identities

(08:30) Policy enforcement and access control for agents

(09:00) The shift toward multi-agent orchestration

(10:03) How different enterprise agents will interact

(11:24) Observability and monitoring agents in production

(13:00) Personal productivity vs enterprise transformation

(15:00) Where companies should start with agentic AI

(18:00) Operating models across IT and business teams

(20:00) Measuring ROI from agents in real deployments

(21:30) Workforce impact and organizational resistance

(23:00) What the next year of agentic AI may bring