Webinar: Add a Banana – The Alchemy of Agentic AI

This webinar, titled “Add a Banana, The Alchemy of Agentic AI,” features host Bob Samuels and guest Chris Penn (Chief Data Scientist at Trust Insights). They discuss the evolution of Generative AI into Agentic AI and its implications for B2B marketing and business operations.


Key Topics Covered:


  • The “Add a Banana” Concept (2:24 – 6:01): Chris Penn explains that AI models are probability engines. To move beyond generic, high-probability “slop,” users should incorporate improbable or unique terms into prompts to force the AI to reconcile different datasets, resulting in more creative and unique output.

  • Agentic AI & The End of Commodity Software (16:07 – 18:55): The discussion shifts toward agentic systems like OpenClaw (or Claude Co-work), which are autonomous systems capable of executing complex, multi-step project plans. Penn argues that because basic knowledge and software are becoming effectively free, B2B companies must move up the value chain from software to transformative experiences (13:06 – 15:36).

  • The Future of Work and Marketing (26:43 – 29:41): The panel explores the impact of AI on employment, noting a shift toward structural unemployment for early-career knowledge workers. Penn suggests that the “hottest programming language” for 2026 is good project management skills, as humans must become effective delegators to AI agents.

  • The Death of SaaS and Billing Models (47:01 – 49:31): Penn highlights the threat AI poses to SaaS companies and hourly service billing (e.g., legal or consulting), predicting that businesses relying on manual, templated tasks will struggle to justify traditional fee structures.

  • Practical Implementation (33:09 – 35:08): Recommendations for AI ecosystems include Gemini for non-technical users in Google environments and Claude for coding-intensive or agentic workflows.

Key Takeaways:


  1. Orchestration over Stacks (54:29 – 55:42): The future is moving away from tool-centric martech stacks to orchestration-centric architectures governed by AI agents connected via APIs.
  2. Human Role: The human touch is not dying but evolving into upfront planning, data curation, and oversight (17:57 – 18:53).
  3. Speed of Change: Technology capabilities are increasing so rapidly that Penn notes prediction is difficult beyond a few weeks (51:24 – 52:17).