Keywords

Agentic AI, AI models, automation, governance, business applications, technology, future of AI, decision-making, AI strategy, customer experience

Summary

In this episode of 'From Ready to Empowered', Duncan Potter explores the emerging field of agentic AI, which refers to systems capable of taking initiative and making decisions. He discusses how these models are built on foundation models like GPT-4 and how they can be applied in various business contexts, such as customer service and marketing. The conversation also delves into the importance of governance in agentic systems, outlining key principles for ensuring their responsible use. Finally, Duncan provides guidance on how organizations can start implementing agentic AI solutions.

Takeaways

• Agentic AI systems can take initiative and set subgoals.
• These systems are built on foundation models like GPT-4.
• Orchestration frameworks enable planning and tool use in agentic AI.
• Agentic AI can analyze job histories and make recommendations.
• In sales, agentic AI can manage outreach and follow-ups.
• Developers are creating agents that can complete projects autonomously.
• Agentic systems are not fully autonomous and need context.
• Governance is crucial for the responsible use of agentic AI.
• Three layers of governance include task boundary clarity, review checkpoints, and context grounding.
• Organizations should start with problems involving multiple steps for agentic AI implementation.