The Elastic Organization

“There is a big debate right now if AI will replace humans or augment them. My view is: it is up to us. So, the best advice I have for you is don’t let AI replace you. To avoid that, become a genuine expert. Then, AI becomes a superpower — you can direct it, challenge it, and do ten times more than you could alone. The students who use AI to avoid thinking will not go far. The ones who use it to amplify their thinking will be extraordinary. The worst thing you can do is let AI replace you in your own education. If the AI is writing your exam, you are training it — not yourself.”

Samuel Huber, CEO MENA at Napster, Founder of Landvault, shared his thoughts on building digital businesses and adding real value with AI to organizations in my "Leading Through Disruption" series.

Lauterbach: Sam, you are a physicist by training. You calculated Lewis Hamilton's engine strategy at Mercedes Formula One — your first job out of university. Why did you leave after only two years?

Huber: I remember the end of my very first day on the job knowing it was not where I would spend my life — even though I had dreamed about exactly this as a kid. I was a small cog in a big machine running against my ambition to be in the driving seat of my own destiny.

Lauterbach: In 2013, you moved into mobile gaming — at a time when Angela Merkel, herself a physicist, was calling the internet “uncharted territory.” What made you see something the German Chancellor did not?

Huber: Genuine curiosity. I am not naturally a gamer, but I caught myself playing simple games on my phone while waiting for meetings. And I thought: if this was happening to me, a non-gamer, millions of others were about to follow.

The best way to build something significant is to catch a wave — if the market is carrying you, you can create real impact. Angela Merkel had decades of legacy to manage. I had nothing to lose by diving into something new.

Lauterbach: You identified a real problem to solve— ad interruptions destroying the player experience — and built your first company around it.

Huber: Every time I integrated ads into a game, the reviews went down. But without ads, the company made no money. Our peers treated ads as a necessary evil. Then I heard Revolut's founder explain how he created his product: International transfers cost twenty dollars not because they should, but because no one challenged it. So, I thought: What if ads were part of the game environment rather than an interruption? Product placements, natively integrated — a real brand's billboard in a racing game, targeted to the player. We proved the model worked.

Lauterbach: You grew the business to 120 people, raised three rounds of funding, then pivoted into metaverse with Landvault. How do you evaluate an emerging trend in real time — and know when to leave the boat?

Huber: The metaverse failed because countless brands jumped in without clear substance and vision. But in the Middle East I found something different. The thinking there was that everything with a physical presence will one day have a digital counterpart — every building, every city, every infrastructure project. That felt real and grounded. I moved to the region, built for clients who needed 3D digital copies of physical environments, and found genuine traction. Each phase gave me the credibility and the capital to enter the next one. My company got acquired for $450m, and through the rollup Napster joint.

Lauterbach: Napster — a brand the music industry tried to destroy. Why keep the name?

Huber: Napster stood for something: the idea that the way things have always been done is not necessarily the way they should be. Specifically, in 1999 Napster tried to break the gatekeepers of music. 26 years later, we are doing the same with AI and data – but the stakes are much higher. When we acquired it, the brand still carried a feeling that the rules can be rewritten. That is exactly what we are trying to do with how humans and organizations interact with AI.

Lauterbach: So what does Napster actually do today? Most companies are investing heavily in AI and struggle with ROI.

Huber: Right now, every AI interaction looks the same: you type a prompt, you wait, you get an answer. That is not how humans communicate. It is not natural, and it is not trustworthy. Our thesis is simple — for AI to be genuinely useful, it needs to feel like a conversation with someone who knows your organization and context. Napster is the Enterprise interface layer between these customers, their employees, and the AI system.

Lauterbach: You use the term “digital employees.” What does that mean in practice?

Huber: Think of it as hiring. Napster has over 20,000 specialist agents — each trained on a domain-specific curriculum: physics, customer service, finance, healthcare. When a company brings one on board, the agent knows the field but not the company yet. The next step is onboarding — exactly like a new human hire. You give them your product documentation, your CRM, your ‘tone of voice’ guidelines. Ninety percent of the value comes from that first-party data integration. After that, the agent can be deployed on various surfaces to interact with your employees or your customers, and learn from them. The insight organizations capture are specific to their business in real time and becomes first party data — not generic advice from the open internet.

Lauterbach: How do you enable AI’s access to real-time information? Most foundational models are frozen after being trained, the knowledge cut-off happens months before deployment.

Huber: The base LLM layer is fixed — you are right. But companies connect their own live data feeds on top of it. So, the agent knows in real time how many products are in the warehouse, what the current pricing is, who is on the team. That combination — general intelligence from training, company-specific knowledge from onboarding, and live operational data from APIs — is what makes a digital employee genuinely useful.

Lauterbach: You described most enterprise AI today as “Frankenstein” — agents bolted on across functions that never ‘speak’ to each other. I encounter this daily in my advisory practice. What is the alternative?

Huber: Treat agents as employees on the org chart, not as software add-ons. In our model, agents know the organizational structure — the reporting lines, the teams, the contacts. If I ask my agent to schedule a call with all US salespeople tomorrow, it does not ask me for a list. It knows who they are, it knows their emails, it sends the message. A marketing agent that does not know what the finance agent is doing might become a bottleneck. We are building an integrated operating model — where human and synthetic employees coexist in the same structure, with shared context.

Lauterbach: I speak to a lot of CHROs and rarely hear a convincing vision of what HR looks like in an AI world. What are you observing?

Huber: In the Middle East, the conversation is completely different. There is a national AI strategy with hard deadlines — fifty percent automation of services within two years. The C-suite is not asking “should we do this?” They are redesigning the entire operating model. Sovereign funds are thinking about what AI-native organizations look like structurally.

In Europe and the US, most HR discussions are still framed around 2015-era talent management. What is missing is the strategic question: given where we want to go, what is this workforce capable of — and what can AI add, and how do we bridge the two?

Lauterbach: You call this the “elastic organization” — borrowed from cloud computing. Cloud made infrastructure elastic. You are proposing the same for labor. What is the biggest bottleneck?

Huber: Culture — not technology. The people in power often feel threatened. AI pilots run for three months, show minimal results, and get stopped because no one will champion the full-scale restructuring. Elastic organizations redesign their operating model to allow humans and agents to collaborate and deliver business results, not just their software stack.

Lauterbach: Final question. What is your advice to a student entering university today — someone who will build their career in the age of AI and robotics?

Huber: There is a big debate right now if AI will replace humans or augment them. My view is: it is up to us. So, the best advice I have for you is don’t let AI replace you. To avoid that, become a genuine expert. Then, AI becomes a superpower — you can direct it, challenge it, and do ten times more than you could alone. The students who use AI to avoid thinking will not go far. The ones who use it to amplify their thinking will be extraordinary. The worst thing you can do is let AI replace you in your own education. If the AI is writing your exam, you are training it — not yourself.