Agentic AI Advances Beyond Hype

Agentic AI is gaining traction as highlighted at the recent Fortune Brainstorm Tech event in Park City, Utah. Microsoft is set to purchase AI from Anthropic, marking a partial shift from OpenAI, while Oracle's stock has surged following announcements of significant AI deals.

Agentic AI refers to autonomous software systems capable of perceiving their environment, making decisions, and acting towards goals with minimal human intervention. Salesforce's Agentforce platform is a key player in this space, and leaders from Zillow, Experian, and Okta discussed the emergence of agentic AI within large enterprises.

Shibani Ahuja from Salesforce introduced a 'maturity model' to understand the evolution of AI agents in enterprises. This model progresses from information retrieval and response to autonomous task handling and orchestrating workflows across multiple domains.

Nicholas Stevens from Zillow explained that their AI agents are transitioning from Level 1 to Level 2, assisting real estate agents with calls, notes, and follow-ups, and gradually taking on more autonomous tasks.

Kathleen Peters from Experian noted that their conversational agents are already helping customers improve credit scores and assisting financial institutions in risk assessment. The future will see increased interaction between agents across different platforms.

Security remains a critical concern. Okta's Bhawna Singh emphasized the importance of protocols like MCP in enhancing AI agent security. Many enterprises are adopting these protocols, which are essential for the advancement of agentic AI.