Scaling a Responsible Agentic Workforce
summary
Claims that 95% of pilots fail and only 5% succeed, prescribes a four-step AI safety system starting with HAX (human AI experience), and requires explicit role definitions, KPIs, ownership, escalation paths and inventorying. Recommends writing governance/model cards in GitHub or Notion, using an escalation-on-threshold human‑on‑the‑loop pattern, and shows a crawl-stage implementation on Marbalism (six agents, promo code team70). Reports running 32 AI employees and stresses zero‑trust access, grounding data, and connectivity as the key success factors.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 3:34Baby tiger analogy for AI
- 7:06First AI employee
- 12:07Why AI pilots fail
- 15:38Roles and success metrics
- 19:11Human-in-the-loop and escalation
- 21:44Inventorying AI agents
- 24:36AI system design mistakes
- 27:27Steps to responsible AI
- 30:58Q&A
related talks
29mThe AI Teammate Is Here — Now What?
56mPanel: Scaling AI successfully
16mLaunching AI Pilots: Insights for Startups and Enterprise
11mMitigating AI Agent Risks: Enterprise Strategies for a Safer Workforce
48mOperationalizing AI Across the Enterprise
28mBuilding an AI-Ready Workforce
26mA reality check on AI’s impact on business transformation
43mFrom Pilots to Practice: Leading the agentic AI transformation