The Agentic AI Masquerade: How to Tell What’s Real vs. Marketing

By Scott Hebner | November 23, 2025

In this episode of The Next Frontiers of AI, Scott Hebner is joined by internationally respected technologist and author David Linthicum for a much-needed reality check on the state of agentic AI and what is really agentic AI.

Scott and David examine the widening gap between rhetoric and reality, drawing on new industry evidence—including the Boston Consulting Group’s finding that only 17% of 1,265 enterprises are actually implementing agentic AI. They unpack the three systemic mistakes fueling this confusion:

Across the industry, companies are loudly proclaiming that they have “AI agents” and “agentic workflows”, yet most of what is marketed as agentic intelligence is little more than repackaged GenAI assistants based on foundational LLMs —prompt chains, deterministic logic, and RAG pipelines dressed up as autonomy. The result is an industry-wide masquerade obscuring what true agentic systems are supposed to be: software that can plan, reason, adapt, and act with measurable independence.

Agentic in name only: rebranding old architectures as “agents.”

Centralized control miscast as autonomy: labeling modular components as agents despite no self-direction.

Misallocated investment: building costly autonomy for use cases that don’t require it.

Together, they define what authentic agentic AI looks like, why definitions matter, and how leaders can distinguish meaningful progress from marketing hype. The conversation provides an unfiltered, pragmatic guide for executives seeking to invest wisely, avoid unnecessary complexity and confusion, and prepare for the “golden age of AI Agents’ as the foundation of digital labor transformation strategies. It’s true agents — not assistants — will reshape digital labor and enterprise workflows, and surely the next frontier of AI. Find out what really is agentic AI.