Data Driven

The conversation focused on the complex realities of deploying AI agents in enterprise environments, moving beyond the hype to examine what these agents should be allowed to do versus what they can do. One concept discussed was the critical importance of defining clear boundaries and authorizations for AI agents—emphasizing that failures often stem from insufficiently thought-out controls rather than the underlying models themselves 02:08.A key theme that emerged was the necessity of human oversight, especially for high-stakes decisions in sectors like healthcare and finance 11:45. The discussion explored evolving practices in agent governance, accountability, and the growing need for transparency as AI systems become increasingly sophisticated and embedded in daily workflows 22:12.Several points were raised, including the risks of over-reliance on AI, the need to maintain auditability, and the consequences of blurring the line between human and AI agency. Set against the backdrop of a rapidly shifting technological landscape, this episode offers practical insights for organizations seeking to harness AI responsibly, while staying alert to new ethical and operational challenges.LinksNitin on LinkedIn -https://www.linkedin.com/in/nitingupta-profile/Watch on YouTube -https://youtu.be/2g_AZqLRib8Time Stamps00:00 Challenges in Enterprise AI Agents04:14 Comparing AI tools: Muse vs. GrokBot08:25 Challenges in AI Adoption and Trust11:45 Establishing AI-human decision boundaries13:18 Automating Agent Supervision18:49 Determining AI vs. human tasks19:49 Automating payments with AI23:13 Human vs. AI Decision-Making27:24 Using AI for presentation feedback32:16 Boosting productivity for small businesses35:20 Ethical questions in tech industries36:35 Impact of AI on Society41:26 Prioritizing tasks for automation43:20 Learning from major tech frameworks

In this episode of Data Driven, Frank La Vigne talks with Richard Luna, CEO of Protected Harbor, about why security built around keeping attackers outside a network leaves organizations exposed once someone gets in. Richard explains his concern that AI-driven attacks will probe multiple weaknesses at once, and makes the case for layered defenses, network segmentation, and infrastructure designed around the applications it supports.The conversation moves from security architecture to the people and processes that make systems dependable. Richard shares how his team plans for failures, works across technical silos, and tests vendor claims before putting equipment into production. Frank and Richard also explore the promise of AI-assisted coding, the experience junior developers need to build, and the expensive gap between a successful demo and software that holds up under real demand.From a storage cluster that failed on its second day of testing to a deployment that collapsed under load, the episode offers practical lessons in building durable systems and teams that take responsibility for them.Key Takeaways• Design beyond the perimeter. Richard argues for barriers and tripwires inside the environment, with attention to how data moves between systems.• Build around application needs. Storage performance, traffic separation, redundancy, and failure recovery should reflect the actual workload.• Break down technical silos. Developers and infrastructure teams need each other’s knowledge to diagnose problems and implement security effectively.• Keep human judgment in AI-assisted development. Generated code and automated tests still need experienced review and realistic load testing.• Make accountability specific. Divide responsibilities clearly, give engineers a path to ask for help, and involve clients when they own a blocker.• Test before trusting. Vendor assurances and a working demo do not establish how a system will behave in production.About the GuestRichard Luna is the CEO of Protected Harbor. His career has spanned software development, managed services, hosting, DevOps, and SaaS infrastructure engineering. In this conversation, he brings that experience to the design, operation, and security of systems that businesses depend on.Links and ResourcesRichard on LinkedIn:https://www.linkedin.com/in/richardluna/Protected Harbor:https://www.protectedharbor.com/Watch on YouTube -https://youtu.be/y6ZHVmJtf9oTime Stamps00:00 Why the 1980s security model falls short01:21 Introducing Richard Luna02:32 What SaaS infrastructure engineering means03:53 AI threats, perimeter defenses, and information silos05:44 Building barriers and tripwires07:17 Application-aware infrastructure10:27 Phased change and lessons from Toyota11:22 Rebuilding infrastructure around the workload13:37 Collaboration, zero trust, and asking for help15:35 Data flow, security, and finding the root cause18:19 Listening to clients and earning long-term trust20:08 AI-assisted coding and the next generation of developers21:17 Why generated tests are not enough25:17 DevOps experience and “works on my machine”26:58 Learning from the loss of a customer29:02 Dividing responsibilities without losing accountability31:56 Oversight meetings that remove obstacles33:34 Empowering clients and engineers35:42 Respect, culture, and learning across teams38:06 Building a business by solving clients’ problems42:01 Planning for failure instead of fighting fires43:01 Stress-testing enterprise storage47:01 A deployment failure and the cost of skipping tests48:32 Cloud scaling and runaway costs49:48 Closing thoughts and connecting with RichardSubscribe to Data Driven for more conversations about data science, AI, data engineering, and the people building the systems behind them.

In this episode of Data Driven, host Frank La Vigne sits down with Andy Pruett, co-founder and CEO of Lumiscent, to explore a topic often overlooked in the industrial sector: it’s not a data problem—it’s a decision problem.Andy explains how industrial companies are drowning in sensor data, but struggle to transform that information into actionable decisions that minimize physical asset risk. The conversation dives into how AI-powered decision intelligence can illuminate hidden vulnerabilities, contextualize risk in real-time, and empower industries like mining, manufacturing, and energy to make smarter, faster calls on the shop floor.From the challenges of aging infrastructure to the growing retirement wave of experienced engineers, this episode reveals how the right blend of technology and expertise is shaping a new era of operational resilience and risk management.LinksAndy on LinkedIn -https://www.linkedin.com/in/andypruett/Watch on YouTube -https://www.youtube.com/watch?v=qzpVpgoO5KALumicent -https://www.lumicent.com/Blog Post -https://www.franksworld.com/2026/09/28/navigating-the-data-tsunami-from-iot-sensors-to-smarter-decisions-in-industrial-operations/Time Stamps00:00 Focusing on asset risk management03:56 Managing food plant risks07:48 Managing information in heavy industries11:45 Workforce challenges and knowledge loss16:00 Transforming data into decisions18:46 Predictive maintenance challenges in plants22:14 Adoption urgency and house analogy27:32 Collaborating with Risk Engineers29:16 Understanding insurance risk strategies32:31 Insurance industry changes and parametrics36:46 Managing risk in operations40:03 Challenges with predictive analytics in cars43:02 Getting into engineering from Alaska45:37 Handling a power outage crisis49:01 Wrapping up and contact info

Welcome back to "Data Driven," the podcast where we explore the cutting edge of data science, AI, and the rapidly evolving world of data governance.In this episode, we sit down with Richa Kaul, founder and CEO of Complyance, an AI-driven platform transforming governance, risk, and compliance for enterprises. Drawing on her experience as a data privacy enthusiast and leader in the compliance space, Richa Kaul dives into the pressing challenges of securing sensitive data, the complexities of global compliance frameworks, and how AI both complicates and protects our interconnected world.We discuss the pivotal role of the Chief Information Security Officer, the importance of proactive, layered security measures, and why compliance is rapidly gaining attention—not just as a regulatory checkbox, but as a central business strategy. With practical advice for both enterprises and individuals on maintaining digital privacy and security, this episode is essential listening for anyone navigating today’s high-stakes data landscape.LinksRicha on LinkedIn -https://www.linkedin.com/in/richa-kaul/Complyance Web site -https://www.complyance.com/Complyance on LinkedIn -https://linkedin.com/company/complyancehq/Watch on YouTube -https://www.youtube.com/watch?v=CFsUcm5u4MYTimestamped overview00:00 Privacy and security frameworks overview03:54 Impact of Equifax data breach06:21 Focusing on data security09:58 Managing complex enterprise security controls15:53 Focus on AI and information security18:37 Learning from Fire Safety Education20:47 Phone security in modern cyber warfare24:48 Concerns about IoT device security29:10 Protecting Privacy from Algorithms30:44 Security and privacy concerns35:23 GTA's social commentary on privacy39:11 AI's Role in Cybersecurity Awareness41:03 Global connectivity and surveillance issues44:47 Ending on a positive note

Welcome back to Data Driven, the podcast where we dive deep into the evolving worlds of data science, business intelligence, and artificial intelligence. In this episode, host Frank La Vigne sits down with Rob Collie, CEO of P3 Adaptive and a former Microsoft engineer, to explore the remarkable journey from Excel’s foundational role in data-driven workplaces to the transformative power of Power BI—and now, to the unfolding impact of AI in business.Rob shares captivating stories from his time on the Excel and Power BI teams at Microsoft, offering insider perspectives on how these tools revolutionized the way organizations work with data. Together, Frank and Rob discuss the critical role of expert stewardship in products like Excel, the complexity hidden beneath familiar interfaces, and the challenges traditional BI faced before the rise of more accessible, user-centric solutions.The conversation then turns to the cutting edge, unpacking what generative AI and large language models mean for the future of data, and why business intelligence might just be the best starting point for companies aiming to embrace AI. Whether you’re a data engineer, AI enthusiast, or business leader, this episode is packed with insights on how mastering data remains the foundation for all technological progress. Join us as we look backward and forward at what it really means to be data driven.LinksRob on LinkedIn -https://www.linkedin.com/in/robcollie/Fair Game: Customizing AI to Your Business Is Easier Than You Think -https://link.amazon/B0bgySLKyGet a free Audiobook on Us -https://www.audible.com/ep/creator?afsrc=1&source_code=PDTGBPD060314004RTime Stamps00:00 Interview with Rob Collie06:38 Managing Excel's core development team08:28 Learning Excel's Complexity12:15 Discovering Power BI at Microsoft15:12 Understanding and using SSAS MDX formula20:23 Making Power BI user-friendly23:29 AI exceeding expectations for crafters26:21 Founding and evolving P3 company28:05 AI model advancements and business impact32:34 LLM memory limitations35:11 Integrating AI into consulting business40:21 Challenges with training AI chatbots41:24 Understanding AI without going deep46:56 AI improves business intelligence48:34 AI enhancing business intelligence51:53 Why LLMs succeed where CPUs fail54:29 Core skills in tech evolution