Agentic AI is not Generative AI (MCP/LAM/LLaVA/CUA vs LLM/NPU/NPG/GraphRAG) feat. Nico Carnaval

Dedicated to Tariq King the King of AI 

As artificial intelligence moves beyond traditional automation and generative capabilities, ensuring AI’s reliability, transparency, and ethical integrity becomes a pressing challenge. This session introduces AI-Assurance, a new paradigm for validating and governing Agentic AI systems, which leverage Computer Use Automation (CUA), Chain of Thought (CoT), and GraphRAG architectures to enhance model-driven decision-making, self-learning, and adaptive agent-based CUA execution.

Building on decades of advancements in cognitive engineering, AI and automation,, this talk explores how large action models (LAM), large vision models (LLaVA) and AI-driven software agents are redefining productivity & assurance across the software development lifecycle (SDLC).

Attendees will gain insights into the evolution of building AI-infused system methodologies, the shift from traditional process automation to autonomous AI agents, and the role of governance models like ISO 42001 and AI risk compliance frameworks (AI TRiSM / ISO / NIST).

Key topics include:

• Agentic AI and the Future of AI-Augmented Workflows – How AI systems move from passive assistance to active moral decision-making.

• GraphRAG and Chain of Thought Reasoning – How structured AI reasoning enhances software quality, transparency, and compliance.

• CUA and Large Action Models – The emergence of autonomous, neuro-symbolic, AI agents for real-time computer use execution.

• AI-Assurance in the Age of Regulation – Navigating compliance, ethical AI governance, and risk management frameworks.

This session is essential for AI practitioners, modern software development, and business decision-makers seeking to integrate AI-Assurance governance into their AI-productivity or AI-workforce workflows while avoiding bans from emerging AI liability laws and EU AI Act compliance mandates.