Bruce Kelley is CTO and SVP of NetScout, leading technology strategies for product and service solutions.

getty

​If you open a junk drawer in your house right now, you’ll probably find a graveyard of old cables: a proprietary Apple cord from five years ago, a Samsung charger from three years ago and a handful of others that don’t fit anything you own today.

For a long time, enterprise software looked exactly like that drawer, with thousands of proprietary “cables” connecting data and tools. If you wanted your security software to feed your operations dashboard, you would hard-code a custom integration. It worked, but it was tedious, expensive and brittle.

The consumer electronics industry eventually admitted that the model was unsustainable. It moved to USB-C, a standard connector that could power almost everything.

Artificial intelligence is pushing a similar shift across technology.

That standard is called the Model Context Protocol, or MCP. You don’t need to understand the engineering details, but you should understand what it represents. It signals a move away from tightly controlled vendor silos and toward a more open, agent-ready ecosystem.

What Is MCP?

MCP defines how AI models discover and interact with tools, data sources and capabilities in a structured way. It’s a framework that allows an AI model to connect to a database, a weather service or a security platform without requiring a custom integration for each.

The impact of MCP will look a lot like the late 1980s, when enterprises began building networks.

The Simple Network Management Protocol, or SNMP, became the universal connector of that era. It standardized the way devices reported information. SNMP led to the development of network management systems, enabling companies to monitor their networks through a “single pane of glass.”

When SNMP entered the market, the number of network agents grew exponentially. That made troubleshooting—human reasoning—possible at scale. In the same vein, MCP is already driving a tsunami of AI agents. Instead of human reasoning, however, it enables AI reasoning.

The Customer Is In The Driver’s Seat

For companies in the networking space, SNMP adoption was a simple matter of life and death. Companies that built walled gardens around their hardware became invisible to network management systems. I’d be surprised to learn that any of those companies survived. Nobody wanted to buy hardware unless it worked with their existing network management tool.

MCP adoption is also a matter of survival. The same is true for companies that build networking solutions—and also for companies that build every other type of software. If the AI can’t see it, it’s as if it doesn’t exist.

Walled gardens are also going to fall across the software industry. For customers, that means they can pick and choose the best providers, assembling a best-of-breed suite of solutions that meets their needs.

The ease of automation, however, will also solve a long list of problems. In many industries, companies face a skills gap as workers retire without replacements. In the near future, I think we’ll see companies use automation to bridge this gap.

Observability In A World Of AI Agents

Perhaps the biggest change MCP introduces is that AI is no longer confined to static models, frozen in time at the moment they were trained. MCP connects AI to real-time data and live systems. What we are talking about is operational AI, with significant implications for security, observability and service quality.

When you move from a monolithic application to a distributed ecosystem of agents, your attack surface expands. If one agent poses a security vulnerability, the whole process is at risk because that agent may touch dozens of other tools.

Likewise, service levels are also implicated. You’ll need to know what tools and data your AI is pulling. Success will depend not only on the quality of your models, but on the quality of your data.

Organizations will need deep, real-time observability into these interactions to ensure that AI-driven decisions are accurate, secure and aligned with business intent. MCP may standardize the connection, like USB-C did for devices, but without clear visibility into what’s flowing through those connections, companies risk plugging into a system they can’t fully see or control.​​

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?