Welcome to the traffic layer.

Kong was built on a bet that software would break into thousands of small pieces, and that every piece would need a safe way to talk to the others.

That conversation happens through APIs. Our co-founders, Augusto Marietti and Marco Palladino, built an open source gateway to sit in the middle of it: one place to secure, route, and watch every request. Developers picked it up fast. Enterprises followed.

Now the same pattern is repeating with AI. Models, agents, and tools all talk to each other through APIs. What changed is what's inside the traffic: prompts, tokens, and data your customers can't afford to lose track of.

Connectivity turned out to be infrastructure twice. First for cloud-native software. Now for AI. That's why we call ourselves The AI Connectivity Company.

Every company you'll call on is pushing AI into production. The numbers on how that's going are sobering.

Speed, cost, and risk. Those are the three things standing between a company and AI that pays off. All three show up in the same place: the traffic. Which models get called. How many tokens get burned. What data leaves the building.

Kong sits in that traffic, so we can do something about all three at once. That's the whole reason we matter.

For the last couple of years, AI mostly lived in pilots. Now boards want it in production, and production changes the questions.

Look at that list again. Nothing on it is about how smart the model is. The projects that stall usually die in the gap between the model and the business: the agent can't reach the data it needs, nobody owns the outcome, and there's no way to pull it back when it goes sideways.

Agents raise the stakes. An agent doesn't wait for a person to click a button. It calls models, APIs, and other agents on its own, all day long. When the UK's AI Security Institute studied more than 177,000 agent tools, the share of usage that takes real action (sending the email, changing the file, moving the money) jumped from 24% to 65% in sixteen months. Every one of those calls is traffic that needs rules.

The companies that put a governance layer in place now will move faster for years. The ones that don't will either slow down or take risks they can't see. Your customers are making that call right now, which means you're showing up at exactly the right time.

Kong's technology is strong. Technology alone doesn't close deals.

The executive who signs off on an AI program isn't thinking about gateways. They're thinking about the board meeting where someone asks how the company is controlling AI risk and spend.

Start there, and you get routed to the people who own outcomes. Start with the product, and you get filed next to every other gateway vendor and handed to a platform team comparing feature lists.

That translation, from business problem to Kong, is the job. Nobody else in the building does it for you.

You don't need to configure anything yet. You do need to explain what Kong does to someone who isn't technical, in about two sentences.

Agents and MCP

MCP, short for Model Context Protocol, is the emerging standard for how agents connect to tools and data. Kong governs that traffic too, so agents get access to what they need and nothing more.

One platform for every kind of traffic

Konnect is where customers manage all of it in one place: API gateways, Kong AI Gateway, event streams, and service mesh. One set of rules, one view, instead of a separate tool for every kind of traffic.