Goodbye dashboards: How MCP Servers are changing the way we work

For years, dashboards have been the standard interface for accessing and understanding business data. Companies invested heavily in building slick UI layers on top of their data. Teams spent time learning which filters to use and where to click. The idea was to put insights into the hands of non-technical staff.

But something new is on the horizon, something that makes dashboards optional.

It’s called the Model Context Protocol, or MCP.

Why should I care about MCP?

If you're like me, when you need some business or project data to make important decisions, you go into a data dashboard. For example, I often open up Produtive to check project budgets. Or I go into our financial forecast Google Sheet to check how much work we  have allocated over the next 6 months. 

This is ok, but I end up fiddling with filters to try and get the answers. And then I have to manually build sheets if I need to correlate data from different places.

However, what if you could do all this securely within ChatGPT or Claude, and let them do all the hard filtering and correlation work? What if you could simply ask questions, and the AI would know how to find the answers using real data?

Suddenly you don't need to understand advanced filtering and querying techniques in your dashboard. You just ask questions and AI will provide answers. In realtime.

Right now, this hasn't been easy to set up. But MCP is a technology that paves the way for it to happen.

The first piece of the puzzle you need to build or plug-in is an MCP server.

What is an MCP server?

An MCP server is a type of AI agent server that understands natural language and uses it to interact with systems, tools, and data on your behalf. It enables you to ask questions or give commands in plain English and get the answers or actions you need.

It knows who you are, what you're working on, and how your organisation operates. In practice, that means you can skip interfaces altogether.

Instead of logging into a BI tool and clicking through filters, you can say:

• "How did our sales compare to last year?"
• "Show me which areas had unusual performance this month."
• "Send the latest operations report to Sarah and copy the regional managers in."

Behind the scenes, the MCP server pulls the right data, formats it, and performs actions based on your request. It acts as both your data analyst and your assistant.

Why this matters

Dashboards were a leap forward. They gave business users access to data without needing to code. But they still require a cognitive load: people need to know what they’re looking for, where to find it, and how to interpret it.

MCPs eliminate those steps. You simply ask.

These servers combine language models, organisational context, secure data access, and task execution into a single agentic interface. The result?

• Speed: No more waiting on dashboards to load or reports to be built.
• Simplicity: Anyone can use them, regardless of their technical ability.
• Context-awareness: They know your role, your current work, and what matters to you.
• Actionability: They don’t just deliver answers, they complete tasks.

What’s happening in the real world?

This isn’t a theory. Big companies are moving fast.

In May 2025, Microsoft announced MCP support in Windows. Their vision is to let users engage with files, applications, and corporate systems using natural language. No UI required. This brings conversational workflows into everyday tools, making agent-based computing part of the OS experience.

Meanwhile, platforms like ChatGPT Teams and Copilot already use similar underlying concepts. They act as a middle layer that bridges user input with organisational data and tools, using secure protocols and contextual awareness.

We're seeing the beginning of a shift where the "app" becomes less relevant. The conversation becomes the interface. 

We’ve even put this to the test for ourselves here at Pocketworks. During our weekly business planning meetings (Rocks), we often find ourselves asking questions about what effect it would have if we were to do X or Y. This usually leads to us having to go and run the relevant reports from the likes of our finance systems, CRM or project management tools and make sense of this data. Instead, we put together an MCP server that had direct access to this information, and now these questions can simply be answered by an LLM using this MCP server.

What does this mean for your business?

For business leaders and product owners, this shift opens new doors:

• Accelerated decision-making: Leaders and teams can access and act on data instantly.
• Reduced tool fatigue: Fewer interfaces mean less complexity and fewer user complaints.
• Smarter automation: Routine tasks like report generation, follow-up emails, or cross-system updates can be triggered through simple instructions.
• Lower training overhead: New staff don’t need to learn a maze of systems; they just ask!

Imagine:

• A CFO says, "Summarise our app performance data for Q3 and draft a board update."
• A product manager says, "List customer feedback themes from the last two months and create requirements for our developers for the top issue."

The MCP server handles the rest.

Where should you start?

You don’t need to go all in right away. But there are smart steps you can take to prepare:

• Assess your data infrastructure: Do your systems have structured APIs or secure access points for these agents to tap into?
• Prioritise clear access rules: These systems need to know who can access what, and under what conditions.
• Start with high-friction workflows: Identify where people waste time switching between tools or requesting data.
• Choose tools with agent readiness: New tools from Microsoft, OpenAI, and others are building native support for MCP-style workflows.

Think of this like the early days of the web or mobile. Those who adopted early often gained years of advantage.

Final thoughts

The way we interact with software is undergoing a quiet revolution.

MCP servers signal a move away from dashboards, search boxes, and complex interfaces. Instead, we’re heading toward a world where you can simply ask, and the system does the work.

It’s faster. It’s simpler. And it’s coming fast.

At Pocketworks, we’re already helping clients explore agent-powered tools and workflows. If you’re curious about how natural language interfaces could reduce complexity and increase agility in your organisation, let’s talk.