Podcast Episode

#379 How to Govern AI Before It Spreads

Embracing Digital Transformation·Dr. Darren Pulsipher·1 September 2026·48 min

About this episode

Check out my new book AI Augmented Teams on Amazon or on my website paidar.ai/books. AI is forcing CEOs to confront a new kind of risk, and Dr. Darren and guest Dennis O'Shea dig into what enterprise leaders need to do before it spreads. From AI governance and data security to Gen Z workarounds, agent management, and AI spend control, this conversation explores why readiness matters more than speed—and how to build an AI strategy that scales safely. ## Key Takeaways - AI doesn’t level the playing field—it exposes weak data, broken workflows, and missing governance. - Most organizations are not ready to deploy AI at scale because use cases aren’t clearly defined. - Data sprawl creates real risk when employees upload sensitive files, emails, or HR documents into public LLMs. - Gen Z is especially likely to bypass friction, making shadow AI and unsanctioned tools a growing governance challenge. - AI rollout works best when leaders classify data, add guardrails, and train frontline workers—not just office staff. - Three emerging enterprise problems to watch: AI spend management, agent lifecycle ownership, and identity/security for AI agents. ## Chapters - 00:00 AI fear, urgency, and why governance matters now - 02:05 Catching up with Dennis: pickleball and AI services - 04:10 Why AI exposes weak processes instead of fixing them - 06:30 The enterprise AI readiness gap and lack of use-case planning - 09:15 Data sprawl, sensitive files, and privacy risk - 13:20 Gen Z, shadow IT, and unsanctioned AI tools - 16:40 Locked-down enterprises and the challenge of secure collaboration - 20:05 Structured AI rollout: data classification and DLP - 23:10 Frontline workers, training, and adoption gaps - 26:15 Mid-market pressure and the role of automation - 30:00 New AI challenges: spend management, agents, and identity - 35:20 The AI-augmented operating system and the book project - 41:00 AI slop, integrity packets, and authentic outputs - 47:10 Using multiple models for critique and validation - 50:00 Where to find the survey and more resources