As GenAI reshapes enterprise automation, many teams have treated prompt engineering as the path to better AI. But prompt engineering alone can't defend against the inherent unpredictability of LLMs. If you're moving AI into production, reliability has to come from architecture, governance, and observability. Not prompt hacks. 👀 Check out this blog from @hguerreroo, to learn why architectural determinism is the foundation for enterprise AI reliability: bit.ly/4vrtKZ8

In the last few months at Kong, we've processed several trillion tokens through our AI gateway. Here's what we're seeing with enterprise AI usage: 1) AI request volume: 56% Anthropic 39% OpenAI 1.5% Gemini (Huge surge from Anthropic over last 6 months to replace OpenAI as the leader) 2) Model preference is not uniform across industries. For example, Claude is dominating in Finance, where the work requires complex, long-context workloads and high governance requirements. 3) The share of AI traffic through the platform has grown more than 6x in the last six months. 4) Average tokens/request is 6,621 and climbing higher as workloads deepen.

Limitations of tool use in AI models. The importance of an AI context layer. SaaS companies becoming an API key. These are just a few of the topics our CEO and Co-Founder, @sonicaghi, spoke about on @MTSlive last week. Check out the full convo 👉 bit.ly/3QDS2A4

SITUATION EXPLAINED: Have the human internet and the agent internet already split apart? We asked @subnetmarco, CTO of @kong "There is already a bifurcation. If we look at our properties right now, 80% of traffic is being driven by agents. It's not even being driven by humans anymore." "Not being able to properly cater to this new agentic traffic means that we're not catering to our users anymore, our customers anymore." "This is not something new. If you were not on the Yellow Pages, your business didn't exist. Then everybody moved to the internet. Then mobile apps. Facebook acquired Instagram because they felt like they were not catering to a whole new buyer moving to mobile." "Likewise, we're seeing that shift yet again, every time there is a shift in the distribution channel, that creates opportunity. It also creates disruption for the ones that cannot follow where the trend has gone." "What determines if one succeeds or fails is the ability to cater to this new buyer, and that is the agent itself."

Introducing the newest member of our Technology partner program 🥁 🥁 @NomaSecurity! Together, we deliver comprehensive agentic AI security by integrating high-speed traffic orchestration with deep, AI-native guardrails. By embedding Noma’s security context directly into the Kong AI Gateway, we provide a hardened execution layer that governs, observes, and protects every agent interaction as it happens. Learn more: bit.ly/4w4rR4v

Our CTO and Co-Founder, @subnetmarco, is live on @MTSlive! Tune in now 👀

MCP: new enough to be exciting, old enough to create a sprawl problem. Tell me if this sounds familiar. One team spins up an MCP server. Then five more. All built differently of course! Nobody knows what's out there or who owns what. Yikes. Agents got flooded with tools they don't need and you accidentally spent a months worth of tokens in a day (and you can't even blame Fable). Shadow infra at it's finest! But now it's got AI speed, so it's like supercharged chaos. That's why we built a central gateway for all AI context. One place to enforce authentication and security for all your data and tools. One point of observability to see which tools are being used, by which agents, and at what cost.

AI is creating a new attack surface. As organizations move AI into production, prompts, agents, models, and APIs are generating a growing volume of interactions that traditional security tools weren't built to protect. The result: fragmented visibility and increased exposure to prompt injection, data leakage, and adversary manipulation. Today, CrowdStrike is extending Falcon AI Detection and Response (AIDR) across a growing ecosystem of AI gateway partners, making the Falcon platform AI's security control plane. Organizations can now: 👁️ Unify visibility across AI activity 🛡️ Detect and stop AI threats in real time ⚙️ Enforce consistent security controls as AI scales Together with @Databricks @GoogleCloud, @jetstream_sec, @Kong, @LiteLLM, @getmaximai, @mspartner, and @TrueFoundry, we're helping organizations secure AI with confidence. Read more: crwdstr.ke/6015BDy2H5

Here's the agent design requirement nobody scoped: 🗣️ "I act autonomously, but I shouldn't bankrupt the company." 84% of companies are reporting AI margin erosion. Most have no token controls, no governance, no audit trail. Don't miss the AI Governance and Token Economics tracks at API + AI Summit Sept 30–Oct 1 in LA. 🦍 Tickets prices go up Aug 15. Register now: bit.ly/4fJebal

📣 Insomnia 13 is here! Insomnia + Kong Konnect now integrate to unify API + AI development workflows. 🧑‍💻 Instantly discover and test any API endpoint 🔄 Endpoints, routes + auth settings are automatically synchronized ⏰ Start testing immediately bit.ly/4xu0UIW

🛠️ We've been working to make Kubernetes the central place to manage more of the Kong ecosystem. Enter: Kong Operator 2.2! It includes Kong Event Gateway support in Kubernetes, Dev Portal support, increased control over the infrastructure Kong Operator creates, and more. Learn about how Kong Operator gives teams a clearer path to standardization and reduces operational gaps between different parts of the platform: bit.ly/4oxjGeH

Don't over-engineer your harness! Lessons learned at @kong from 34,000 tests using Agentic AI. 👇 konghq.com/blog/engineering/…

One of the best ways to avoid model lock-in is to use a multi-LLM AI Gateway that allows to seamlessly send requests to multiple models without having to rewrite your code. At @kong we built an enterprise solution that allows to achieve just that, via the universal API. Build once, switch to any model instantaneously.

The biggest obstacle to cloud migration isn't technology. It's complexity. 😵‍💫 Many organizations know they need to modernize, but the risk of moving critical API infrastructure keeps them stuck on legacy platforms. 🤝 Together with @Persistentsys, we help enterprises migrate with confidence while building the connectivity, governance, and security foundation needed for AI. Learn more in our blog post: bit.ly/4vK11yA

Kicking off Day 1 at #FOST Amsterdam! Great sessions, great conversations, and a dedicated @AsyncAPISpec track focused on the future of event-driven architectures. If you're here, stop by and say hi to the @Kong team. We'd love to chat. #EventDrivenArchitecture #Kong

We've been hard at work building out the agenda for API + AI Summit. Here's what you can expect. 👇 5️⃣ tracks: AI governance, API modernization, event-driven agentic AI, AI monetization + cost control, and context engineering. Plus all of the hands-on workshops and networking we can fit into 2 days. Early bird pricing ends tomorrow! Register now > bit.ly/3QA8GAt

From Melbourne to Stockholm to Washington DC, the Kong Agentic Era World Tour is off to a roaring start! 🦍 If you're an engineer or technical leader who wants to learn what it takes to win in the age of AI connectivity, this is for you. We'll have business tracks on governance and cost control, technical workshops, and roundtable discussions with other decision-makers building AI in production. 🌎 The World Tour ends this month. Check out the remaining stops here: bit.ly/4e9TotR

Great week at #APIops Helsinki! Thanks to Marjukka for the invitation and to everyone who attended, shared experiences, and joined the discussions. Proud to represent Kong and exchange ideas on API governance, platform engineering, AI, and developer experience. #APIs #Kong

AI agents can tell you what they know. But can they tell you how they got there? 🤔 AI observability needs needs more than prompts, tokens and model responses. Developers need a complete, replayable history of every decision, tool call, and context change. Check out this blog from @hguerreroo on why durable commit logs may be the missing foundation for AI observability 👉🏼 bit.ly/4v37Xqv