Insights on MCP and AI connectivity with Ray Bogman, Head of Innovation, Alumio
Developed in 2024 by Anthropic, creators of the Claude LLM, the Model Context Protocol or MCP is rapidly emerging as a foundational standard for connecting AI assistants and LLM tools with thousands of business applications and data sources. MCP provides a standardized, model-agnostic interface that any AI assistant or tool can use to interact with a wide variety of applications and datasets, without creating bespoke integrations. As such, it's being described as the USB-C for AI.
To understand the real-world applications of MCP better, we interviewed our resident AI expert and Head of Innovation, Ray Bogman, to give us a more grounded perspective on this new universal standard for AI connectivity. Here’s what he had to say:
What is MCP, and how is it impacting AI integrations?
“MCP, or Model Context Protocol, is like a plug-and-play solution that standardizes the way AI models, especially LLMs, can retrieve and augment data from external applications and data sources. It’s very similar to what we do with the Alumio iPaaS— bridging systems and data. The impact of MCP in the world of AI has been quite significant, since Anthropic introduced it in 2024. It doesn’t just impact how AI functions; it changes how integration is approached entirely. It’s not just an AI advancement—it’s a new way of thinking about connectivity.”
How can Alumio iPaaS users effectively leverage MCP?
“One of the key advantages of MCP is how it allows AI models to directly interact with business systems in a context-aware way. With Alumio acting as the integration layer, this becomes incredibly powerful for our customers.
For example, let’s take our customers who use the Alumio iPaaS to connect to Magento (now known as Adobe Commerce). In the near future, we envision offering an MCP server connection specifically for Magento within the Alumio ecosystem. What this means is that an AI assistant, through MCP, could query data from Magento using simple prompts. For instance, a user might ask, “What T-shirts are available in the summer collection?” or “Which SKUs belong to our ‘essentials’ category?”. With MCP in place, those prompts can trigger real-time data retrieval from Magento via Alumio, without the need for custom queries or additional configurations.
This turns what used to be complex data lookups into natural, AI-driven interactions. Alumio sees it as an opportunity to enable our customers to use AI not just for content or automation, but to dynamically augment their product catalogs, order systems, or any connected app, directly from within Alumio. It’s a smarter, more flexible way to use AI across the entire integration landscape.”
What are some other use cases for MCP?
“There are quite a few promising use cases of MCP, the potential of each depending on the availability of existing Connectors and how an application exposes its data. While Magento (Adobe Commerce) is one example, we also work with applications like Spryker, Shopify, Shopware, BigCommerce, and many others, each of which could support MCP-enabled connections through the Alumio iPaaS.
The beauty of MCP is that it builds on top of what’s already there. For instance, with Magento, we’re not asking customers to modify their systems or expose new endpoints. We’re simply using the standard REST APIs that already exist. Alumio helps configure all the intelligence and logic, such as the context formatting and model communication, which is required for the AI to understand and access the data via MCP. In other words, by leveraging basic APIs, it gives customers the ability to query the data that’s already connected to their Alumio routes.
That means the same principle can apply across other platforms too. If a business is using Shopify, for example, they could prompt an AI assistant to pull real-time stock levels, fetch order history, or even generate a sales summary, without needing to create new integrations. It’s about unlocking access to existing business data in a way that’s prompt-ready and AI-compatible.
Ultimately, MCP allows Alumio to serve as the bridge, so any system we connect to today via APIs can become AI-accessible tomorrow, with minimal friction. That’s what makes it so scalable.”
How does MCP enable AI assistants and agentic workflows?
“There’s a strong and growing relationship between MCP and AI assistants, often referred to as agents. These agents are basically AI-powered autonomous decision-makers or orchestrators. You give them a prompt or a goal, and they figure out how to get it done by reaching out to the right tools, services, or data sources.
MCP acts as the gateway that allows these agents to access the real-time data or tools they need to complete a task, without needing to build new custom integrations. So if you ask an agent, “What were my top-selling categories today?”, MCP allows it to go fetch that directly from your system, using the APIs that are already there.
And this is just the beginning. Currently, most agents handle one task at a time. But what we’re moving toward is something called agentic workflows, where multiple agents collaborate, automate tasks, and even make small decisions on their own. Imagine one agent pulling in stock data daily, another checking trends, and a third creating a sales report, all running quietly in the background.
In that sense, MCP isn’t just another connectivity method—it has the potential to transform how organizations design, implement, and scale integrations in an AI-native world. It gives AI agents a common language and a safe way to interact with business systems. It’s about creating an ecosystem where you start to delegate real business processes to AI.”
Is MCP the future of AI-driven integrations, or just another connectivity layer?
“Over the years, integration protocols have evolved significantly, starting from early technologies like FTP and XML-RPC to more modern standards such as SSH, SFTP, SOAP, REST APIs, GraphQL, gRPC, and Webhooks. Each introduced new capabilities, improving how systems connect and exchange data. MCP (Model Context Protocol) represents the next major shift in this evolution—designed specifically for AI.
What sets MCP apart is that it’s this golden midpoint that brings AI into the integration layer, not just for connecting systems, but for augmenting how we interact with them. Unlike traditional protocols that simply move data, MCP enables AI models to access and interact with that data contextually. In other words, until now, we could only ask AI models questions based on what it had already been trained on. With MCP, we can now ask those same questions, but to any sort of real data, from live systems, where you can type in prompts to get data or answers from applications in real-time.
We’re entering a world where any data source, application, or service can become part of a broader AI-driven system. While current frameworks like RAG (Retrieval Augmented Generation) pull in data from one source, MCP allows us to access multiple sources. Every system can act as a server or a client, where we can use AI to dynamically query your connected data in real time. So yes, I believe MCP has the potential to reshape the way integrations work entirely.”
Preparing for MCP as the next big shift in AI and integrations
What MCP reveals isn’t just a technical innovation, but a shift in how we approach integration in an AI-driven world. It signals a move away from building one-off connections or manually querying APIs, toward a model where AI can navigate, retrieve, and act on data from across systems, using the integrations businesses already have in place.
For businesses building and managing these integrations with solutions like the Alumio iPaaS, this opens up a new layer of value. Existing connections become more than integration routes, they become access points for intelligent, prompt-based interactions. And as MCP becomes more widely adopted, the groundwork customers have laid through Alumio, positions them to experiment with AI safely, without the complexity of reinventing their architecture.
It also reframes the future role of integration itself. With MCP, integration won’t just be about syncing data between systems anymore. It will be about enabling context-aware, real-time dialogue between AI and the tools businesses use. That makes AI more usable, more practical, and more impactful, across teams. And as organizations begin delegating more tasks to AI, protocols like MCP may become the glue that holds the complex agentic workflows that arise together.
It’s not just about what AI can do - it’s about how seamlessly it can fit into the systems we’ve already built”.
Ray Bogman
Head of Innovation, Alumio
