THE TRUTH ABOUT AI

Why the Agentic Enterprise will define the next decade of the AI revolution

- BY MARC BENIOFF

IN HAMPSHIRE AND THAMES VALLEY IN SOUTHERN England, a new kind of cop is on the beat—and it doesn't wear a uniform. “Bobbi,” built on my company’s Agentforce platform, is an autonomous system that handles non-emergency calls to the police—the thousands of routine questions each day that can overwhelm response centers and slow response times. Bobbi now addresses many of these calls instantly, freeing human officers to focus on the critical emergencies that require, with lives hanging in the balance, the judgment, empathy, and on-the-ground experience that only humans can offer.

Bobbi reflects an important shift in how AI is showing up in the world. What makes the platform work isn’t just a large language model (LLM) like ChatGPT or Claude. It’s the way AI is woven into existing law-enforcement data, connected to the systems and applications police already rely on every day, and guided by people who set priorities and intervene when needed. That’s what I call the Agentic Enterprise: AI, data, apps, and people all working together as one unified system that turns intelligence into action.

I’VE BEEN PART of many waves of technology through my career, from the cloud to mobile, social, and now AI. Each arrived with enormous excitement, and sweeping assumptions. These assumptions—like the belief that social media would bring the world closer together—rarely played out the way people first imagined. True to form, with AI we are now moving past the honeymoon phase of dazzling expectations and into real-world complexity.