From an Emmental childhood and elite junior hockey to Microsoft cloud, international speaking, enterprise AI and robotics.

I will always remember the morning I unexpectedly ran into my future co-founder in the city. Christof Schnyder was carrying a backpack containing the CHF 20,000 in capital required to establish our GmbH, moving it from one bank to another.

In that moment, our freedom to shape the future became real. I did not yet understand all the challenges it would bring, but the values behind the company — responsibility, resilience, curiosity and service — had formed much earlier.

This is the story of how a young ice-hockey goalkeeper from the Emmental discovered technology, became a Digital Workplace Architect, built a Swiss Microsoft agency and eventually started a new chapter in AI engineering.

Before technology, there was hockey

I grew up with my parents and sister in Switzerland's Emmental, where the original Emmental cheese is made.

It was a grounded childhood in a rural region where community, reliability and hard work matter. At ten, I began playing ice hockey as a goalkeeper for SC Langnau.

At sixteen, I left the Emmental and moved to Bern to pursue elite junior sport.

A goalkeeper develops a particular view of the world. You see the entire system in front of you. You learn to anticipate what might happen next, communicate under pressure and remain calm when everything starts moving quickly.

You also learn responsibility.

When the puck crosses the line, there is nowhere to hide. You analyse what happened, reset and prepare for the next shot.

I did not know it then, but those lessons would later shape how I approached customers, technology, leadership and entrepreneurship.

Discovering technology at eighteen

At eighteen, I discovered computing and almost immediately began turning that curiosity into a business.

In 2001, I founded comm.across ITC, my first PC-support business. I worked directly with small and medium-sized companies, solving everyday technology problems.

There was no distance between a technical decision and its consequences. If a system stopped working, someone could not do their job. If I built something badly, I had to return and fix it.

Technology should help customers become more successful at what they do.

That principle still gets me out of bed today.

In 2004, I founded Webkor Websolutions, designing websites and custom business applications for small businesses.

During the same period, I joined isolutions. I started as a Solution Engineer and later became a Technical Consultant. My entrepreneurial experience with smaller customers began to meet the complexity of larger enterprise systems.

I was learning that good technology requires more than technical ability. It requires listening, translation and an understanding of how organisations really operate.

A friendship that started in 2006

Microsoft collaboration technology, better known as SharePoint, and I first met in 2006.

Those were the days when SharePoint appeared to be the centre of the enterprise universe. Mike Fitzmaurice took the stage in his Scottish kilt. SharePoint Designer was preparing to replace FrontPage. Portals, document libraries and collaboration sites felt like the future of work.

Looking back, it was a wonderful chapter in my career.

SharePoint taught me that enterprise technology is never just about software. A digital workplace touches communication, knowledge, processes, identity, governance and organisational culture.

A technically correct platform can still fail if people do not understand it, trust it or see how it helps them.

That lesson has never changed.

Following Microsoft into the cloud

In 2009, I moved into Microsoft's cloud with BPOS, the Business Productivity Online Suite and predecessor to Office 365.

The journey continued through Office 365, Microsoft 365, Teams, OneDrive, Power Platform, Azure AI and Copilot.

SharePoint never really disappeared. It became quieter and more specialised, operating beneath modern intranets, Teams, OneDrive and countless business processes.

Moving to the cloud also transformed my role. I was no longer thinking only about servers and applications. I had to think about continuously evolving services, identity, information architecture, security, adoption and governance across entire organisations.

In 2012, I received the Microsoft Community Contributor Award for my contributions to the Office 365 forums.

The award mattered because it recognised something I still consider fundamental: sharing knowledge.

Technology communities succeed when people share what they have learned. One answer, documented failure or clear explanation can help someone else move forward.

Knowledge becomes more valuable when it moves.

Building a Microsoft business on moving ground

Building a company around Microsoft technology has meant building on ground that never stops moving.

In the on-premises era, customers controlled the servers, upgrade cycles and much of the surrounding infrastructure. That control came with its own burden: complex migrations, accumulated customisations, hardware dependencies and major version changes that could become projects lasting months or years.

Then hybrid IT forced us to connect two worlds. Identity, data, security and governance crossed the boundary between customer infrastructure and cloud services. The old environment did not disappear when the new one arrived.

BPOS and Office 365 changed the model again. Customers gained speed and shed infrastructure responsibility, while agencies adapted to an evergreen service whose features, interfaces, licensing and practices could change during active projects.

Microsoft 365 expanded both the opportunity and the complexity. SharePoint became part of a larger system spanning Teams, OneDrive, Power Platform, security, compliance, identity and data. Mastering one product was no longer enough. We had to translate a constantly changing environment into stable customer decisions.

Now AI is compressing the cycle once more. Models, tools and development methods can change in weeks rather than years. Customers must make decisions about data boundaries, security, governance, cost and human accountability while the technology itself is still developing.

Staying relevant requires more than collecting certifications. It takes energy, curiosity and a willingness to question what you know: learning while delivering, investing before demand is certain and letting go of expertise that once defined your value.

You also have to protect customers from two opposite mistakes: moving too slowly because change feels uncomfortable, or moving too quickly because a new technology is surrounded by hype.

That balance has been one of the hardest parts of building CNEXT. Our responsibility is not simply to stay on top of every announcement. It is to understand what has materially changed, what remains true and what customers should do next.

Platforms keep changing. The obligation to create durable customer value does not.

The beginning of CNEXT

In 2012, I founded my first limited company under the Collaboration NEXT name.

It represented what I believed the next generation of enterprise collaboration should become: connected, cloud-based, human-centred and continuously evolving.

In 2014, Christof Schnyder and I established a new legal entity together. The Collaboration NEXT name was transferred to the company we would build jointly.

Our original idea was deliberately modest. We wanted to work as independent senior professionals: experienced people doing meaningful work directly with customers, without building a traditional agency hierarchy around us.

Reality had other plans.

Only one year later, we were already four people. From there, we grew slowly and intentionally. We never wanted CNEXT to become large for the sake of appearing successful. We wanted to preserve seniority, trust, direct customer contact and the freedom to do good work.

During the coronavirus years, we peaked at 15 people on the payroll. On paper, that looked like progress. In practice, it changed the company.

More of my energy went into management and maintaining the organisation. It became harder to stay close to customers, technology and the craft. Fifteen people were too many for the company we intended to build.

Today, CNEXT is back to around ten people. That feels like the right size: large enough to bring together complementary skills and deliver substantial work, but small enough to remain personal, senior and accountable.

This journey changed how I think about growth. Bigger is not automatically better. A healthy company is not defined by the largest possible headcount. It is defined by whether its size supports its purpose, its people and the quality of its work.

My own role evolved with the organisation. I moved from Technical Consultant to CTO and co-founder, and eventually to CEO and Solution Architect.

The titles changed, but the purpose remained close to the customer.

I wanted to understand how organisations operated, where information was lost, which processes created unnecessary friction and where technology could make people more effective.

CNEXT was never only about deploying Microsoft products. It was about making those products useful.

Becoming a Digital Workplace Architect

I came to describe my role as a Digital Workplace Architect. To me, architecture is not drawing diagrams from a distance. It is understanding what people need to accomplish, where information gets lost and where technology can remove friction without adding another interface.

My focus expanded across SharePoint, Microsoft 365, Teams, OneDrive, Power Platform, Azure AI, Copilot and intelligent automation. The platforms changed, but the mission remained the same: create intelligent, connected workplaces that customers can operate and improve themselves.

Giving power to the people

As enterprise computing moved into the cloud, I specialised in Microsoft's citizen-development technologies. From 2018, Power Apps, Power Automate and Power BI became central to my work. I later became certified as a Microsoft Power Platform Solution Architect Expert.

Power Platform let people close to a business problem participate directly in solving it. Sometimes our citizen developers challenged the "real developers." Sometimes we even won.

The goal was never competition. Professional engineering, security and governance remain essential. The goal was inclusion: giving more people the power to turn ideas into useful systems. AI is now accelerating the same movement.

Learning in public across borders

Community has always been part of my career.

Between 2017 and 2019, I completed 24 documented speaking appearances across Europe and New York. I spoke about SharePoint, citizen development, intelligent workplaces, bots, low-code business applications and applied AI.

Several sessions were delivered with Jussi Mori, including The Rise of the Citizen Developer at Microsoft's Times Square facilities in New York and our applied-AI session at Microsoft AI Bootcamp Vienna.

Jussi and I believed technical sessions should be useful, honest and fun. A live demonstration should occasionally make your heart beat faster.

Speaking internationally taught me to explain complex systems clearly, challenge my assumptions and understand how organisations in different countries approached similar problems.

Teaching made me a better architect. Community feedback made me a better founder.

My complete public speaking record

• 2017: The Rise of the Citizen Developer at SPS Milan, Paris and Lisbon; SharePoint Communication Sites: Deep Dive at Microsoft 365 User Group Zurich.
• 2018: The Rise of the Citizen Developer at SPS Madrid, Warsaw, Netherlands, New York City, Stockholm and Lisbon, SharePoint & SQL Saturday Cambridge, SPS Vienna Preconference and Thrive Conference Ljubljana; SharePoint Communication Sites: Deep Dive at SPS Zurich; Low-Code Intelligent Business Applications in Office 365 at Microsoft AI Bootcamp Vienna; How to Get Started with Modern SharePoint for HappitChat; and a Hyperfish company-directory session at SPS Helsinki.
• 2019: The Intelligence Infusion in Modern SharePoint at SPS Brno; The Intelligence Infusion in SharePoint at SPS London; How Intelligent Is Your Digital Workplace? at aOS Moscow and Modern Workplace Conference Paris; and How Intelligent Is Your Modern Workplace? at SPS Central Europe, Stockholm and Leicester.
• 2026: Top Builders with the Replit Community.

This record matters because it shows that my work on intelligent systems did not begin with the generative-AI boom. Citizen development, conversational systems, workplace intelligence and applied AI were already central to my thinking years earlier.

Turning customer pain into products

Building CNEXT taught me that useful products begin with repeated customer pain, not brainstorming workshops.

Storage Monster grew from seeing organisations pay for Microsoft 365 storage without clearly understanding its consumption. We turned that operational and financial problem into a product that makes waste visible and optimisation actionable.

The founder lesson was simple: listen carefully, find recurring friction, make it visible and build with the people who experience it every day.

What robotics taught me

In the Swisslog environment, I led engineering around the technical foundation supporting high-availability AutoStore operations.

The work connected warehouse management, conveyors, picking systems, real-time computing, monitoring, automation and international service. These systems had to operate continuously.

When physical operations depend on your technology, "mostly working" is not acceptable. Uptime, observability, incident response, documentation, training and operational handover are part of the system, not aftercare.

Production exposes every weak assumption.

A robot does not care how impressive the presentation was. A warehouse cannot pause because the prototype needs another sprint.

The system either works within its operating environment, or it does not.

From virtual assistants to embodied AI

My path into AI did not begin with ChatGPT.

We were already building virtual assistants with Microsoft Bot Framework, Cognitive Services, LUIS, automation and enterprise workplace systems.

The technology was more limited than today's models, but the central question was already clear: How do we connect intelligence to real organisational context without losing control?

The CNEXT AI Labs team I co-lead later developed a project using the open Misty social-robotics platform. The team connected Azure Speech, Azure OpenAI, video analysis, automation, APIs and live ERP data to a physical robot.

That brought together enterprise context, conversational interaction and embodied technology. First, we helped people find and share information. Then we automated processes. Software learned to participate in conversations. Now AI can use tools, create applications and increasingly interact with the physical world.

I see one continuous story: reducing the distance between human intention and useful action.

Agentic engineering changed the speed

Generative AI and software agents have changed what a small, experienced team can accomplish.

I now use agentic environments such as Replit and GitHub Copilot across research, architecture, coding, testing, evaluation and documentation.

Becoming a Replit Level 5 Master Builder reflects how deeply this new way of working has entered my engineering practice.

A senior engineer working with capable agents can explore and deliver at a speed that previously required a much larger structure.

But speed is not responsibility.

AI can generate code, but it cannot accept accountability for a production failure. It can propose an architecture, but it does not own the business consequences. It can accelerate execution, but it cannot decide which risks an organisation should accept.

That is why I call the emerging discipline AI-infused engineering.

AI strengthens experienced engineers. It does not remove the need for judgment, accountability or operational knowledge.

Bear Intelligence and the next chapter

Today, my own company is Bear Intelligence GmbH.

MH – AI Labs is its dedicated forward-deployed AI engineering offering and the next chapter of my founder journey.

Most organisations can create an impressive AI demonstration. The difficult work begins when it must use real identities, respect data boundaries, connect to operational systems, remain economical and fail safely.

Traditional delivery divides that work across strategy, suppliers, platform teams, security and business users. AI exposes every gap between them. MH – AI Labs was created to remove that distance.

We embed AI-infused senior engineers directly into the customer's operational context. We stay close to the business problem, the people, the data and the infrastructure.

When a mission requires specialists in security, cloud, data, hardware, robotics, law or change, we assemble a focused team around it with shared responsibility from problem to operable system.

Architecture should follow the constraint

I have spent much of my career working with Microsoft technology. I understand the value of enterprise platforms, integrated identity and managed cloud services.

But no vendor should determine an organisation's architecture by default.

The right AI system may run in a public cloud, private infrastructure, local hardware, an edge device or a physical machine. It may combine commercial platforms with open-source components.

The customer's requirements should determine the architecture, not the supplier's preferences.

This is not ideology. It is engineering.

The future will leave the browser

Today, enterprise AI is dominated by copilots, chat interfaces and software agents.

That is only the beginning.

Intelligence will increasingly move into sensors, facilities, vehicles, machines and personal robots. The boundary between digital and physical systems will continue to disappear.

Personal robotics is therefore part of the long-term direction of MH – AI Labs.

It remains an emerging practice, not a mature service claim. The responsible work today is readiness: architecture, edge inference, data boundaries, supervision, integration and safety.

Before autonomy can be trusted, the surrounding operating model must be engineered.

The future deserves ambition, but it also deserves honesty.

Fourteen years later

Fourteen years of CNEXT, 20 years of SharePoint and 17 years of Microsoft cloud have not changed the reason I do this work.

I still get up in the morning to help customers become more successful.

The platforms will continue to change. SharePoint gave way to the cloud. Citizen development expanded who could create software. AI agents are transforming engineering. Robotics will bring intelligence further into the physical world.

But the essential questions remain:

• Does this technology solve a meaningful problem?
• Can people use it?
• Can the organisation trust and operate it?
• Does it leave the customer stronger and more independent?

I did not create MH – AI Labs to produce more AI theatre.

I created it to shorten the distance between an idea and an operable system: from cloud to private infrastructure, from software agents to physical machines, and from human intention to useful action.

The goalkeeper from the Emmental is still looking at the complete system, anticipating what comes next and preparing for the next shot.

Technology is the instrument. Human success is still the mission.

For more information, visit Marcel's personal website. Feel free to reach out directly at any time.