Kong

AI agents won’t enter the enterprise through a pile of API keys. With Kong AI Gateway, MCP access now runs through enterprise identity: central policy, real audit trails, and EMA/ID-JAG brokering for clients and servers that aren’t ready yet. Enterprises shouldn’t have to choose between moving fast with MCP and waiting for every client and server to catch up. More at https://lnkd.in/dfX2Vi3M

Today we added a large enterprise customer scaing 1 trillion tokens per month. Agents and models are devouring APIs at machine speed. The internet is becoming the nervous system of digital intelligence — Kong is building the railroads (dendrites) for it. P.S. If you drive by the 101, look up!

⚡️ KAi v2 is live! ⚡️ In v2, KAi, the Kong Konnect agentic platform assistant, goes from a knowledgable guide to something that can actually build inside your organization with guardrails that you control. With our new Konnect MCP Server, KAi can create, update, and delete resources directly in Konnect. It handles creating the service, route, plugins, and consumer, before handing it off to you. What else is new? 🔹 Get downloadable artifacts in the format you need 🔹 Generate full analytics dashboards based on what matters to you 🔹 Increased visibility into consumption with session and daily use widgets 🔹 Admins always stay in control, disable write operations if needed Check out the demo video below to see how the Konnect MCP Server works. 👇 Then, learn more about the new functionalities in this blog from Alex Drag: https://bit.ly/4vtvQqR

Back in February, I started talking about a pattern that I believed would become essential for enterprise AI adoption: the MCP Registry. At the time, the conversation was mostly about building MCP servers and connecting agents to tools. But my view was that at scale, enterprises would inevitably need discovery, governance, lifecycle management, and control over those MCP assets. Fast forward a few months, and at KubeCon Europe, I noticed a clear trend. Session after session touched on the same challenge from different perspectives: how do we manage an ever-growing ecosystem of tools and MCP servers without creating operational chaos? That's why I was excited to read Gartner's new 𝗜𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻 𝗜𝗻𝘀𝗶𝗴𝗵𝘁: 𝗠𝗖𝗣 𝗥𝗲𝗴𝗶𝘀𝘁𝗿𝘆 report (https://buff.ly/SxhgNlK) from analysts Shrey Pasricha, Keith Guttridge, and Andrew Humphreys. One quote perfectly captures why this pattern matters: "𝘔𝘊𝘗 𝘙𝘦𝘨𝘪𝘴𝘵𝘳𝘺 𝘪𝘴 𝘯𝘰𝘵 𝘢 𝘤𝘰𝘯𝘷𝘦𝘯𝘪𝘦𝘯𝘤𝘦 𝘧𝘦𝘢𝘵𝘶𝘳𝘦. 𝘐𝘵 𝘪𝘴 𝘢 𝘤𝘰𝘳𝘦 𝘪𝘯𝘧𝘳𝘢𝘴𝘵𝘳𝘶𝘤𝘵𝘶𝘳𝘦 𝘤𝘰𝘮𝘱𝘰𝘯𝘦𝘯𝘵 𝘢𝘴 𝘱𝘢𝘳𝘵 𝘰𝘧 𝘺𝘰𝘶𝘳 𝘣𝘳𝘰𝘢𝘥𝘦𝘳 𝘈𝘐/𝘔𝘊𝘗 𝘨𝘢𝘵𝘦𝘸𝘢𝘺 𝘪𝘮𝘱𝘭𝘦𝘮𝘦𝘯𝘵𝘢𝘵𝘪𝘰𝘯." It's always rewarding when a pattern you've been advocating for starts becoming industry guidance. This is exactly why we didn't just write blog posts about it at Kong —we built it. We integrated these enterprise-grade MCP patterns right into Kong Konnect and Kong AI Gateway because your AI agents shouldn't be blindly talking to unmanaged data sources without a strict control plane enforcing discovery, authentication, and routing. I believe we're reaching the point where MCP Registries will be viewed the same way we view API gateways and service registries today: foundational infrastructure for building scalable, governable, and secure AI systems. #AI #AgenticAI #MCP #ModelContextProtocol #AIInfrastructure #PlatformEngineering #APIManagement #KubeCon

Gartner just released a new Magic Quadrant, and it's forcing the industry to answer a tough question: What does AI governance actually mean? In the new MQ, Gartner formalized AI governance as a distinct enterprise buying category, and they project the market will be worth $1.4 trillion by 2030. But we have to be precise about what this category does and does NOT include. As outlined by Gartner, AI governance platforms are built for CISOs, compliance officers, legal teams, and risk functions. Their job is to manage things like dynamic risk scoring and compliance framework mapping (EU AI Act, NIST AI RMF, ISO 42001). This is the "what" part of AI governance. But it doesn't cover the "how". Gartner is explicit about this point: governance platforms do NOT enforce policy in isolation. They depend on something beneath them to make those decisions operational at runtime. That's the "how" layer, where Kong lives. Applying AI governance at the traffic layer. Rate limiting, access controls, prompt inspection, PII sanitization, content filtering, etc. This is the enforcement infra that makes governance decisions scalable. It's like traffic law vs traffic lights. You can set broad policies, but you need the traffic layer enforcement to make it actually work. A policy that says "no PII crosses this boundary" does nothing until something in the request path actually checks and enforces it. So what is AI governance? It depends on who you are. CISOs can focus on the "what" layer, while builders need to obsess over the "how". Orgs have to treat governance and AI connectivity as complementary infrastructure decisions. One layer defines the rules. The other makes them real. Traffic law AND traffic lights.

Token spend becoming a problem? An AI gateway at the traffic layer gives you 3 levers that can drastically reduce token consumption. 1) Prompt Compression: strips unnecessary characters from a prompt before it ever reaches the foundation model. 2) Semantic Caching: caches responses based on meaning, not exact wording, so duplicate intent doesn't trigger a redundant model call. 3) Semantic Routing: routes prompts to lower-cost models based on intent, reserving expensive models for complex tasks and cheaper ones for simple requests. These are valuable at any size org, but at enterprise scale (millions of daily requests) the token cost optimization could reshape your budget entirely.

𝗪𝗲 𝘀𝗽𝗲𝗻𝘁 𝘆𝗲𝗮𝗿𝘀 𝗯𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗼𝗯𝘀𝗲𝗿𝘃𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝗳𝗼𝗿 𝗺𝗶𝗰𝗿𝗼𝘀𝗲𝗿𝘃𝗶𝗰𝗲𝘀. 𝗪𝗲 𝗵𝗮𝘃𝗲 𝗯𝗮𝗿𝗲𝗹𝘆 𝘀𝘁𝗮𝗿𝘁𝗲𝗱 𝗯𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗼𝗯𝘀𝗲𝗿𝘃𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝗳𝗼𝗿 𝗮𝗴𝗲𝗻𝘁𝘀. Over the past few months, I've had the opportunity to speak at and attend several industry events focused on the intersection of Event-Driven Architecture and AI. The excitement around Agentic AI is incredible, but one recurring theme keeps coming up in conversations with architects and platform engineers: 𝗛𝗼𝘄 𝗱𝗼 𝘄𝗲 𝗮𝗰𝘁𝘂𝗮𝗹𝗹𝘆 𝗼𝗽𝗲𝗿𝗮𝘁𝗲 𝘁𝗵𝗲𝘀𝗲 𝘀𝘆𝘀𝘁𝗲𝗺𝘀 𝗶𝗻 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻? Too many teams are trying to fix broken agent behavior by playing "prompt whack-a-mole." 🛑 An agent hallucinates, takes an unexpected action, or reaches the wrong conclusion, so we tweak the prompt, rerun it, and hope for a better outcome. That's not debugging. That's guessing. One of my biggest takeaways from these conversations is this: 𝗦𝘁𝗮𝘁𝗲 𝘁𝗲𝗹𝗹𝘀 𝘆𝗼𝘂 𝘄𝗵𝗮𝘁 𝗶𝘀. 𝗛𝗶𝘀𝘁𝗼𝗿𝘆 𝘁𝗲𝗹𝗹𝘀 𝘆𝗼𝘂 𝗵𝗼𝘄 𝗶𝘁 𝗯𝗲𝗰𝗮𝗺𝗲 𝘀𝗼. For agentic systems, history matters. If an agent made a decision, called a tool, or changed its context, we need a durable, sequential record of those events. Because in production, and especially in regulated environments, "the model said so" will never survive an audit. We need a permanent reasoning trace. That's why I've been exploring the idea of using a durable commit log as a foundation for AI observability and why I believe event-driven patterns have an important role to play in the future of agentic systems. I dive deeper into these ideas in my latest article for Kong. 👇 Read it here: https://buff.ly/E1jdbGF #AIAgents #AgenticAI #GenerativeAI #EDA #EventDrivenArchitecture #Observability #ApacheKafka #SoftwareArchitecture

Did you catch our CTO and Co-Founder, Marco Palladino, on MTS last week? Check out the clip below to hear his take on how there is already a bifurcation of a human-facing internet and an agent-facing internet. 🤖 Then, watch the full episode to hear more from Marco on the rising agentic economy, how engineering autonomous loops are replacing manual prompt writing, and more. Link below. ⬇️