On this episode of the Scouting For Growth podcast, Sabine VdL talks to Marinela Profi, Global Market Strategy Lead for AI, GenAI and Agentic AI at SAS, about the rise of agentic AI and how we will move from hype to real, reliable AI.

In today’s episode, we’ll discuss: Why LLMs alone don’t solve business problems – and what does, how governance is becoming the new frontier of AI trust, and what leaders should expect by 2026, as enterprises shift from experiments to autonomous, explainable intelligence.

KEY TAKEAWAYS

A generative AI chatbot is really good and answering questions, generating text, or summarising content. But, it typically stops when it comes to conversation. On the other hand, an AI agent goes beyond that, it can take action, it has goals, memory, reasoning capabilities and can orchestrate multi-set workflows using a combination of not just large-language models but also rules, data and analytics. Generative AI talks, and agentic AI does.
The 5-step lifecycle of an agent is a framework I put together to help me and my customers understand what an agent actually does step-by-step in practice. 1. Perception 2. Cognition 3. Decisioning 4. Action, and 5. Learning.

Governance boards in 2026 will act more like digital oversight committees, they will ensure that agents aren’t just smart, but they are safe, explainable and accountable.

BEST MOMENTS

‘Post action the agent learns from feedback from a human operative. It’s important to monitor the learning loops, you cannot allow the agent to “self-update” in ways that are uncontrolled.’

‘How autonomous should an agent be? 90% of the time it depends on the risk and impact of the task.’

‘Autonomy without accountability is a risk multiplier.’

‘Governance doesn’t stop at deployment, performance must be continuously monitored.’

ABOUT THE GUEST

Marinela Profi helps organizations move from AI hype to trusted impact. As Global Market Strategy Lead for AI, GenAI and Agentic AI at SAS, she works with enterprises in financial services, healthcare, and government to build AI systems that don’t just act fast—but act responsibly.

With an MBA and a Master’s in Statistics and AI, Marinela bridges two worlds: translating complex data science into clear business strategy. Her work focuses on how agentic AI—intelligent systems that perceive, reason, and act autonomously—can deliver governed, explainable decisions instead of black-box predictions.

A frequent keynote speaker at international AI and analytics events, she shares insights on the evolution from generative to agentic AI and the new frontier of AI governance, trust, and human-AI collaboration. Marinela is also an Advisory Board Member for Wake Technical Community College’s Data Science Program, helping shape future-ready curricula that connect classroom learning with real-world AI innovation.

ABOUT THE HOST

Sabine is a corporate strategist turned entrepreneur. She is the CEO and Managing Partner of Alchemy Crew a venture lab that accelerates the curation, validation, & commercialization of new tech business models. Sabine is renowned within the insurance sector for building some of the most renowned tech startup accelerators around the world working with over 30 corporate insurers, accelerated over 100 startup ventures. Sabine is the co-editor of the bestseller The INSURTECH Book, a top 50 Women in Tech, a FinTech and InsurTech Influencer, an investor & multi-award winner.

This Podcast has been brought to you by Disruptive Media. https://disruptivemedia.co.uk/

The War on the Mind (memory corruption, schemas, & neuropsychology of trauma) - 082

The Deep End w/Taylor Welch

Why Taylor Will Write a Book on Mental Warfare

The AI-powered Podcast Player