Have You Heard of MCP? Understanding the Game-Changing Protocol Transforming How Businesses Use AI

Author: Sean Harris - Kyva CEO

‍

Yesterday, I asked Kyva to compare copper prices to the S&P 500. Within seconds, the AI generated multiple tables showing closing prices and percentage changes across different timeframes, complete with an analysis of correlation patterns and potential leading indicators. All this happened seamlessly, without me having to specify data sources or write a single line of code.

Spoiler alert: I'm not about to get rich trading this correlation.

How did it access this financial data so effortlessly? The answer lies in a breakthrough called Model Context Protocol (MCP), which I'm thrilled to announce we've integrated into Kyva.

What is MCP in plain English?

Think of MCP as a universal translator between AI models and the digital world around them. Before MCP, getting AI to interact with data sources, tools, or specific formatting requirements was like trying to have a conversation through multiple translators—clunky and error-prone.

MCP creates a standardized way for AI models to understand exactly:

• What tools they have access to
• How to properly format their responses
• What data sources they can pull from
• What actions they're allowed to take

In my copper vs. S&P 500 example, behind the scenes, the AI used MCP to connect directly to a financial data API, pull the historical pricing information, and process it—all without me needing to explain how to access that data.

Understanding APIs: The Digital Connectors

To appreciate why MCP matters, it helps to understand APIs (Application Programming Interfaces). In simple terms, APIs are like waiters in a restaurant:

You (the customer) don't need to know how the kitchen works to order food. You just tell the waiter what you want, and they bring it to you.

Similarly, APIs allow different software systems to communicate without needing to understand each other's internal workings. They're how weather apps get forecast data, travel sites check flight prices, and banking apps access your account information.

Before MCP, connecting AI to these APIs required custom code for each integration. It was like hiring a different translator for each restaurant you visited.

What This Unlocks: Real-World Examples

With MCP, you can now use natural language to access information across your business systems:

1. "Find all emails from client ABC Corporation discussing the Q3 deliverable timeline." - MCP enables AI to search your email system and return relevant communications without complex search syntax.
2. "What were our total marketing expenses last quarter compared to budget?" - AI connects to your financial systems to pull actual vs. budgeted figures and present a clear comparison.
3. "Which client projects are currently behind schedule and by how many days?" - Your AI queries project management tools to identify delayed projects with specific metrics.
4. "Who on the development team has experience with both Python and AWS?" - MCP allows AI to search employee skills databases and cross-reference multiple criteria.
5. "Show me our top 5 customers by revenue this year versus last year." - AI connects to your CRM, retrieves and ranks customer data, and presents comparative analysis.
6. "Which warehouse locations currently have stock of product XYZ-100?" - Your AI checks inventory management systems across locations and returns availability information.
7. "Find all policies that need updating based on new industry regulations." - MCP enables AI to identify documents with outdated regulatory references that require revision.

Why This Matters for Business

This isn't just a technical improvement—it's a fundamental shift in how we can use AI:

1. No more prompt engineering gymnastics: Instead of crafting elaborate instructions hoping the AI understands, MCP provides clear guardrails.
2. Consistent, structured outputs: Get reports, tables, and analyses formatted exactly how you need them, every time.
3. Tool use without coding: AI can now seamlessly access databases, APIs, and other tools without custom integration work.
4. Reliability at scale: When deployed across an organization, every team member gets the same high-quality experience.

The era of unpredictable AI outputs is ending. With MCP, we're entering a world where AI consistently delivers exactly what you need, in the format you need it, with access to the tools and data required to get the job done.

Want to harness these AI capabilities without exposing your proprietary data to external model providers? At Kyva, we've built a solution that keeps your data within your control. Connect with me to learn more

‍