For the last two years, corporate AI strategy has largely revolved around "Assistants", chatbots that wait for a human prompt, but could that phase be already coming to an end?

Nicholas Eayrs of Databricks argues we are now entering the era of the Agentic Enterprise.

In this new paradigm, AI stops waiting for instructions and starts interacting with other systems autonomously. Imagine a "Procurement Agent" negotiating with a "Finance Agent" to optimise supply chains in real-time, without human intervention.

We explore how this shift from "chatting" to "doing" is already happening, and much more, including: 

• The Death of the Prompt: Why the future isn't about better prompt engineering, but about architecting autonomous flows where humans act as supervisors, not operators.

• The "Multi-Agent" Swarm: How specialised, smaller models (rather than one giant LLM) will collaborate to handle complex tasks like finance and logistics.

• Bridging the Data Gap: The role of the Model Context Protocol (MCP) in allowing AI agents to "talk" to legacy systems of record (like CRMs and ERPs).

• Solving "Black Box" Anxiety: Why a bad chatbot answer is annoying, but a bad agentic decision is expensive, and how Evaluation Frameworks and strict data lineage provide the necessary governance.

The Human in the Loop: Why the modern workforce must pivot from "users" of software to "Systems Architects" who define the constraints and outcomes for autonomous agents.

Produced by: Roshan Kanesan

Presented by: Roshan Kanesan

This and more than 60,000 other podcasts in your hand. Download the all new BFM mobile app.

Categories:  technology,  Corporates,  managing,  investments

Tags:  invisible AI,  agentic ai,  gen ai artificial intelligence,  robotics automation,  Data governance,  LLM,