The Physics of Reinvention: How KPN is Using AI to Transform Its Culture, Not Just Its Tech
At CASA25, Miruna Anastasoaie, a particle physicist turned telco transformation leader at KPN, delivered one of the most refreshing keynotes of the event—not because of dazzling AI demos or grand future visions, but because of her brutally honest, human, and deeply insightful take on what really matters when a telco wants to transform with AI: culture, mindset, and leadership.
If you thought transformation was about plugging in tools or writing slick AI strategies, think again.
From Physics to Telco: A Mindset That Matters
Miruna didn’t come to KPN to maintain the status quo. With a background in astrophysics and a career spanning academia, banking, and telco, she’s seen how hard it is to change large systems from the inside. But she also carries a scientist’s mindset: clear end goals, tolerance for the unknown, and a commitment to iteration.
“In physics, you might know your destination and your first step—but you must relearn at every step. That’s the mindset I brought to KPN.”
And she didn’t come because she thought telco was glamorous. Quite the opposite.
She assumed KPN was too old, too well established, with too much history, a company with deep heritage and complex legacy systems. But when a former colleague pitched her the transformation journey underway, she thought:
“If we can change an operator like KPN, we can do anything.”
The Real Challenge of AI Isn’t Tech—It’s Letting Go
KPN isn’t lacking in tech. It has one of the most complex IT landscapes Miruna has ever seen. But that complexity comes with history—and emotion.
Many of the systems in place have been serving the company for 20 or 30 years and were built by people who have grown with KPN during that time. Letting go of those “babies” isn’t just a technical task—it’s personal.
“That’s why, even as a tech nerd, I focus on the cultural side of transformation.”
Changing mindsets is harder than upgrading tools. Because tools can be bought. But to use them well, you need clarity of purpose, leadership courage, and the ability to unlearn.
The AI Illusion: It’s Not About Tools
Miruna challenged a popular misconception head-on:
“Is this really an AI problem? Or is it a mindset problem?”
She described meetings where everyone was focused on “getting the best LLM,” without understanding that an LLM without retrieval-augmented generation (RAG) is just a parrot.
The obsession with new tools misses the point. No tool solves your core issues unless your foundations are solid:
- Clean data
- Clear processes
- Purpose-driven leadership
And perhaps most importantly:
“No supplier can solve our problems for us. Partnerships are important but you still own your problems. And ultimately you solve them from the inside.”
From Use Cases to Scale: Stop Planting Flowers
One of the biggest laughs—and biggest truths—came when Miruna spoke about the focus on use cases when she joined:
“It was like everyone was planting their own flower. But AI needs to work at scale, not in silos.”
Instead of spreading dozens of use cases across departments, she emphasized building shared journeys, connected processes, and interoperable AI systems that can grow with the business.
Use cases aren’t the goal—they’re just the first step. And they shouldn’t distract from the hard work of unification.
Don’t Freeze. Don’t Jump. Walk Forward With Purpose.
Many companies either freeze when faced with AI uncertainty—or jump headfirst into implementation with no structure.
Both are dangerous.
Some people thought they weren’t ready and wanted to cancel. She asked them, “What do we have?” When they listed it out, it was enough to move forward.
Miruna shared how her own team faced this when preparing for a board meeting:
“Delimit the known from the unknown. That’s what makes you strong.”
This mindset—that clarity trumps perfection—is what she calls the “physics of reinvention.” Take the first step. Reassess. Iterate.
Agentic AI Needs Orchestration—Not Chaos
When asked about agentic AI, she warned about another common trap: deploying too many agents too fast without guardrails.
“Ten agents solving a broken coffee machine may just argue with each other.”
Building agentic systems is like building teams—you need clear roles, coordination, and governance.
Throwing more energy at a system doesn’t help if the direction is off. In fact, it creates waste. Smart AI implementation is thoughtful, slow at first, and deeply structured.
Identity First. AI Second.
Perhaps the most powerful moment came near the end:
“Who are you, really? What’s your identity as a company? AI should help you—not define you.”
Too many AI projects start from the tech and forget the soul of the business. KPN, under Miruna’s leadership, is flipping that script.
She spends her time educating, challenging assumptions, and asking the tough cultural questions:
What is KPN’s purpose? What do we want to become? What are we willing to change—and what should remain?
Final Thought: The Hard Work Is Worth It
Transformation isn’t linear. It’s messy. It’s full of false starts and difficult conversations. But as Miruna reminded us:
“If you want impact, you have to do things differently. Not because you know better—but because you’re not afraid to fail and learn.”
And that’s what real AI transformation looks like.
Not plugging in a tool—but shifting a mindset.
Not buying intelligence—but building it—together.
