Move Fast And Don't Break Things: Secure AI Adoption with Samantha Mehta

🎙️ In this episode of Beginner’s Guide to AI, Dietmar Fischer talks with Samantha Mehta, solutions engineering leader at AIRIA, about how companies can adopt AI without losing control. If your teams are already experimenting with ChatGPT and AI tools, the real question is not “Should we use AI?” but “How do we use it safely, visibly, and profitably?” Samantha explains what enterprise AI security looks like in real life, including AI guardrails that can audit, block, redact, and replace sensitive data. She also unpacks AI governance and AI observability, because you cannot manage what you cannot see. A key theme is shadow AI and AI sprawl: people will use AI anyway, so organizations need sanctioned paths that reduce risk while accelerating adoption. On the practical side, this conversation goes deep on agentic workflows. Samantha describes how agents become more than prompts through routing, actions, approvals, looping over documents like CSVs, and scheduled runs that create repeatable outcomes. From internal GPT alternatives to workflows that touch expenses, supply chain planning, and customer support, the episode is packed with grounded examples and a clear starting path. 📧💌📧 Tune in to get my thoughts and all episodes, don't forget to ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠subscribe to our Newsletter⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠: ⁠⁠⁠⁠ beginnersguide.nl⁠⁠⁠⁠ 📧💌📧 About Dietmar Fischer: Dietmar is a podcaster and AI marketer from Berlin. If you want to know how to get your AI or your digital marketing going, just contact him at argoberlin.com Chapters 00:00 Welcome and why Samantha got into AI 01:26 What ARIA does: build, test, secure, deliver enterprise AI 02:19 Real use cases from simple internal GPT to complex workflows 08:27 How to start: guardrails first, then build your first agent 11:32 Agentic workflows explained: routing, actions, human in the loop 17:12 Why security and governance matter and why blocking fails 31:14 AI sprawl and shadow AI: monitoring and risk management 40:00 Wow use cases and the future: Blade Runner, change, and jobs 48:42 Where to find Samantha and ARIA Quotes from the Episode 🪧 “I personally can’t think of a case where an LLM needs to know my social security number.” 🪧 “People are going to use it no matter what. If you don’t enable safe usage, they’ll still use it.” 🪧 “Agentic workflows are so much more than just ping an LLM and get a response.” 🪧 “I always say: build, test, secure, and deliver your usage of AI.” Where to find Samantha: ➡️ LinkedIn: Samantha Mehta on LinkedIn ➡️ Company: look at what AIRIA does Music credit: "Modern Situations" by Unicorn Heads Hosted on Acast.