Why the model context protocol is not just an IT project

The next revolution in computing is quietly reshaping how we interact with our machines through a concept called the Model Context Protocol, or MCP. It effectively connects LLM chatbots like Claude and ChatGPT with any existing software or tool. 

But to treat this as just another IT project would be a critical error for businesses hoping to remain competitive.

The stark reality is that most companies that are aware of the fundamental shift that MCP represents are approaching MCP entirely wrong.

MCP was first released in November 2024; you don't need permission from the tool creators to create an MCP server of your choice. So the open source community has been hard at work creating MCP servers for major tools like Hubspot, Notion, AirTable etc

When the graphical user interface emerged in the 1980s, it democratized computing by replacing cryptic command lines with intuitive visual metaphors. 

MCP represents a similar paradigm shift. Rather than humans learning to communicate in machine language, 

MCP enables AI systems to understand human contexts—industry-specific knowledge, unwritten company processes, and the subtle ways expertise manifests in different domains.

But there's a fundamental misunderstanding happening in boardrooms.

AI itself is in danger of being relegated to IT departments, treated as merely another technical implementation. This approach fundamentally misses the point.

Soon, the user interface we're all familiar with where employees log in and interact with the software the company has decided for them, will disappear. 

Instead, at login will be a simple chatbot with the ability to connect with not only any piece of information on the internet or any company database, but also any piece of software in the world or indeed to create any piece of software for the employees' needs.

The difference won't be technical competence. It will be context and personal choice.

Traditional IT departments excel at system implementation, security protocols, and technical integration. 

These skills remain vital but insufficient. The primary value of MCP isn't technical—it's personal. It allows all employees with access to AI systems to choose their tool stack, their own way of work and the differentiation will be in the critical thinking and domain expertise required to build a unique tech stack that works to their advantage. 

In my work implementing AI across various industries, the pattern is unmistakable: when business leaders treat AI as merely technical infrastructure, they achieve technically sound implementations that fail to deliver business value.

But MCP is the antithesis of this. It runs on the individual users' machines, and the MCP tools and workflows they decide to implement are unique to them. 

This isn't to diminish IT departments—they remain essential partners. But leadership must come from those who understand the business contexts being encoded.

Successful implementation requires asking different questions. Rather than "How do we implement this technology?" business leaders should ask "How will our employees use this technology for themselves? What can we learn from them? For retail, this might be context-aware customer service. For healthcare, it could involve clinical decision support that understands practice variations.

The competitive implications are significant. Organisations treating MCP as business transformation rather than technical deployment and employee-led transformation will create systems that understand their specific contexts—a proprietary advantage that competitors can't easily replicate.

The most successful implementations I've witnessed share a common approach: they begin with awareness on an employee level and they are - at their core, creative.

As an AI consultant,I've seen firsthand that implementation starts with awareness and knowledge—the use cases that employees come up with themselves are what make the business unique and AI implementation successful. 

The MCP revolution isn't primarily about technology—it's about preparing for a new world; a world where software and tools are employee-led, through natural language and not top-down SaaS subscriptions organised by the IT departments.

Businesses that understand the potential of MCP and AI and reimagine their business processes around it will be the ones to succeed in the 2020s and beyond. 

And that transformation requires leadership that extends far beyond the server room.