Feb 22, 2026
Episode 11
Scaling AI isn’t as simple as flipping a switch. In Episode 11, we uncover what it really takes to prepare for agentic automation in business.
Agentic automation is rapidly moving from experimentation to board-level priority, but scaling it safely requires more than advanced models. Enterprises must first establish strong data governance, standardize fragmented data environments, and control unstructured data before autonomy is introduced into core workflows. At the same time, human-in-the-loop AI and clear accountability mechanisms are essential to ensure transparency, auditability, and regulatory alignment. For C-Suite leaders, enterprise AI readiness is no longer about launching pilots—it is about building the governance, compliance, and operating foundations that allow intelligent systems to deliver measurable value without increasing risk.
Agentic automation refers to AI that can autonomously interpret information, choose actions and execute multi‑step work, operating as a self‑directing system within defined guardrails. It transforms automation from rule‑based task execution into intelligent, goal‑driven operations.
It represents the next evolution in enterprise automation—AI systems that can independently plan, decide and execute work across entire workflows, not just individual tasks. Unlike traditional RPA, which follows predefined rules, or copilots that support humans interactively, agentic systems operate with goals rather than instructions, enabling them to perceive context, adapt to changing conditions and deliver outcomes autonomously. This shift from task automation to outcome automation fundamentally reshapes how organizations design processes, govern AI and structure accountability.
For many companies, the conversation around AI has moved from curiosity to urgency. Experiments are everywhere, proofs of concept are piling up, and leadership teams feel the pressure to keep up. Yet beneath the excitement, there’s a growing realization: the leap from isolated pilots to agentic automation is far bigger—and more demanding—than most expected. Episode 11 of The Power of Possibility, featuring Christian Schierjott, Head of Digital Transformation and Innovation at SPS, and Bertram Weiss, Vice President of Health at Merantix Momentum, dives into this tension and explores what it really takes to make AI work at scale.
Despite the hype, most organizations remain at the early stages of true readiness. Legacy systems, fragmented processes and inconsistent data still shape the day‑to‑day reality inside many enterprises. Add to that tightening regulation and the rising complexity of AI models, and it becomes clear why the “just plug it in” mindset falls short. Agentic automation doesn’t sit on top of the business; it threads through it, touching every layer of how work gets done.
The episode offers a kind of reality check—not to slow companies down, but to help them direct their energy where it matters. The central question isn’t “How do we adopt AI?” but “How do we build an organization where AI can actually operate meaningfully and safely?” It’s a shift from chasing tools to preparing the ground.
One of the strongest takeaways from the discussion is that agentic automation depends far more on foundations than on models. Clean, accessible, reliable data isn’t a technical nice‑to‑have—it’s the difference between a system that behaves predictably and one that can’t be trusted. Many businesses still struggle with scattered databases, conflicting information and processes that have never been fully aligned. When an AI system has to navigate all of that, the cracks become visible immediately.
But even with good data, companies face another pressure point: governance. Regulations in Europe, Switzerland and the U.S. are rapidly evolving, and autonomy without oversight is no longer an acceptable approach—especially in industries handling sensitive information or high‑stakes decisions. As the episode stresses, human oversight remains essential not because AI “fails,” but because organizations need traceability, auditability and control.
Instead of viewing governance as a brake, the conversation frames it as an enabler. Companies that invest early in clear guidelines, monitoring and responsibility frameworks are the same ones that will be able to scale confidently later.
The biggest challenge we still face is the limited explainability of the algorithms. If I don’t start with the right data, I make it even worse to interpret the outcome
Christian Schierjott
In agentic systems, the interplay between data quality and governance is what ultimately determines whether an organization can scale safely and effectively.
A particularly striking part of the episode explores the European dilemma. On one hand, Europe has some of the most advanced ethical standards and regulatory frameworks. On the other, this strength often comes with a cultural mindset that makes experimentation slower and risk‑taking less common. Many organizations want every uncertainty addressed before taking the first step—which, ironically, increases the risk of falling behind.
The biggest problem actually I see also in the culture is the temptation to wait and see… We may end up as a museum of the old economy, while the US and China are moving forward.Bertram Weiss
The episode warns that waiting for the “perfect moment” to adopt AI may become Europe’s most expensive choice. While the U.S. and China continue to move at extraordinary speed, prioritizing value over perfection, European companies risk becoming mere users—not creators—of the next generation of AI technologies.
Failures in AI initiatives rarely stem from the algorithm. More often, they emerge from the combination of weak foundations: outdated infrastructure, low‑quality data, governance gaps, unclear ownership or simply a lack of executive backing. When these issues stack up, even promising pilots stall.
What the episode makes clear is that companies need to focus on several internal steps before aiming for automation at scale:
None of these steps are glamorous, but together they form the groundwork for sustainable progress.
When things fail, we see a concatenation of errors across technology, culture and organization… In combination, they really make the project stall.– Bertram Weiss
As the episode points out, AI isn’t something an organization “buys”—it’s something it grows into. Those who make the effort now will be the ones positioned to unlock the real value of agentic automation when it matters most.
Preparing an organization for agentic automation isn’t about adopting the latest model—it’s about getting the fundamentals right. As highlighted throughout the conversation, the real differentiators are data quality, cross‑department integration, strong AI compliance, and a human‑in‑the‑loop framework that preserves trust while enabling scale. Most companies don’t fail because the technology isn’t ready; they fail because their processes, culture, or governance aren’t. Those that invest early in solid foundations, build a balanced portfolio of use cases, and commit to structured change management will be the ones able to move from pilots to production with confidence and measurable impact.
Talk to SPS about designing an enterprise AI governance framework and human-in-the-loop operating model for safe agentic automation.