Kong joins ServiceNow Control Tower Alliance aimed at helping enterprises safely adopt AI. As a founding member, Kong will help bring enforcement to the APIs, models, and tools that agents rely on. Kong Inc., the AI Connectivity Company, announced it is a founding member of the ServiceNow Control Tower Alliance, a new ecosystem of leading companies in AI, cloud infrastructure, security, governance, value measurement and token metering, who will work with ServiceNow to help customers adopt trusted enterprise AI safely at scale. The alliance is built around the ServiceNow Super Intelligence Control Tower, a centralized control plane to discover, secure, govern, observe, and measure enterprise AI. As part of the alliance, Kong will work alongside ServiceNow and the other founding members to help bring security and governance to the models, tools, and agents that AI applications depend on. "AI governance depends on enforcing what agents can access and do," said Augusto Marietti, CEO and co-founder of Kong. "Together, ServiceNow and Kong can turn enterprise AI policies into control points across APIs, models, and tools, helping customers put AI agents to work with confidence." Oct 29, 2026 Prev Next 1 of 43,824 As a founding member of the Control Tower Alliance, Kong will work with ServiceNow in three ways: * Build together: Kong and ServiceNow will collaborate on solutions that run on the Super Intelligence Control Tower and are easy for customers to deploy, bringing Kong's AI traffic governance to enterprises inside the platform they already use to govern AI. * Deliver with governance: Kong's solutions will go through ServiceNow's qualification and certification process, ensuring alignment against shared standards for AI governance, so customers adopt with confidence. * Sell and lead together: Kong and ServiceNow will create joint solution offerings that are deeply integrated, offering additional value to customers and partners. Kong enables a universal enforcement layer for enterprise AI through Kong Konnect, the AI Connectivity Platform. It gives enterprises consistent enforcement of security and access policies across the APIs, models, and tools their agents use, with the freedom to change models, providers, and environments without rebuilding governance. Within the alliance, Kong brings this enforcement to LLM, MCP, and agent-to-agent traffic through Kong AI Gateway, translating enterprise AI governance into real-time control over how AI connects to the business. [To share your insights with AiThority, please write to [email protected]]

Kong announces Context Management for efficient AI agent use. BySD Times October 14, 2026 Kong today has announced Kong Context Management, designed to ensure agents can access more of the enterprise context it needs without swamping the model with everything it could use. The new capability is generally available today as part of Kong AI Management. Large enterprise APIs can have hundreds of operations, which when exposed through MCP, turns into hundreds of tools, Kong noted in its announcement. So, even if the model only needs one of those tools, it may need to process all of them, and this "becomes a problem of context consumption, token cost, latency and tool selection," Kong wrote in its blog announcement. That, Kong wrote, creates four compounding problems: Context window tax. Every tool definition consumes context that could otherwise be used for the user's request, conversation history, retrieved information, and the model's reasoning. Reasoning degradation. The agent has to choose among an increasingly large collection of tools, often containing similar or overlapping operations. More choices can make selecting the right tool harder. Higher token costs. Tool definitions consume tokens whenever they're sent to the model. Unnecessary definitions can become a recurring cost across every agent interaction. Higher latency. More context to process and repeated model-to-tool interactions can translate into slower agent experiences. To ensure the agent can access the capabilities behind the API, Kong's Code Mode presents a smaller interface for agents to find and execute the operations they need at runtime, Kong said. An agent can search the available API surface for the operation it needs, retrieve the relevant schema on demand, and execute it without requiring every available operation to be represented as a tool in the model's context. With Code Mode, agents can use code to compose multiple operations, including logic such as loops, conditionals, and error handling. That can reduce the number of model round-trips required to complete complex work while keeping unnecessary tool definitions out of the model's context. The result is an architecture designed to improve three things that matter when agents move into production: * Cost: Reduce unnecessary token consumption and model interactions. * Performance: Reduce context overhead and unnecessary round trips. * Quality: Give the model a more focused set of capabilities to reason over at any given time.

Why MuleSoft, not Agentforce, is Salesforce's most strategic AI asset. Oct 14 2026 MuleSoft Agent Fabric is MuleSoft's recent enhancement to deliver an agentic AI control plane. During the Dreamforce keynote, I noticed something subtle but significant. I didn't see "MuleSoft Agent Fabric." Instead, I saw "Agent Fabric," with the MuleSoft branding dropped and the Salesforce logo replacing the Mule logo in the UI. Forrester sees agentic AI consisting of three planes: the build plane, the orchestration plane for runtime, and the control plane for governance. Long term, I believe the first two will converge: Where you build an AI agent often dictates where you can run it. But it will be impossible to shoehorn all your agents into one build-and-orchestrate platform, resulting in multiplatform sprawl for that part of the agentic AI stack. In contrast, the need for one central place to govern AI agents will push agentic AI control planes to be multiplatform - just like how API management governs APIs across any number of API runtime platforms. That makes the control plane more strategic than the other two. For Salesforce, its emerging agentic AI control plane is Agent Fabric. AI governance is the future of API Management. Leading API management vendors recognize that the agentic AI control plane is the future of the market. I'm even inclined to rename the market "API and AI management" for my next Forrester Wave(TM) evaluation of API management. MuleSoft and competitors Kong and WSO2 have built strong AI gateways to govern LLM and MCP connections. But like APIs, governance is more than just a gateway. Gravitee, Kong, MuleSoft, and WSO2 deliver deep analytics for MCP and LLM token FinOps. MuleSoft has expanded its API portal to include LLMs, MCP servers, and agent cards. Gravitee's beta portal is slated to deliver similar capabilities soon. Both vendors are the first to deliver endpoint versioning for LLM, MCP, and A2A endpoints. Forrester clients can explore this more through Forrester's vendor comparison tool for API management vendors, based on my evaluative research, The Forrester Wave(TM): API Management Software, Q3 2026. Salesforce's advantage over API Management vendors: the full platform. But there are two things that MuleSoft, or rather Salesforce, has to augment its agentic AI governance that those other API management vendors lack: a broad agentic AI orchestration platform and context data to feed agents at runtime. To be clear, it is not assured that Salesforce will deliver a viable AI orchestration + context platform. Today, it is an amalgamation of Agentforce, Data 360, Headless Toolkit, AIforce, Informatica, MuleSoft's iPaaS for deterministic workflows, Slack, and more. Bolted-together Frankensteins create friction through unduly complex user experiences and complicated SKU purchases. But at least it's something. This could go in one of two paths. The ideal path is that Salesforce unifies its Frankenstein into a holistic AI platform with minimal buying and UX friction under one brand. One can imagine a future where Salesforce's agentic AI offering consolidates into three products - and I mean three real products engineered as such, not marketecture. The more likely path is that Salesforce stumbles to deliver. The obvious problem is that bringing together disparate M&A products into something that works is difficult. Notice in the above figure, though, that the control plane is the only layer that isn't currently a stitched-together Frankenstein. Unlike the rest of the platform, Salesforce's AI agent control plane is already one holistic product: Agent Fabric. Between its strategic positioning for buyers, its holistic product design, and its ability to extend Salesforce's reach to agents in every platform, MuleSoft's Agent Fabric, not Agentforce, is Salesforce's most strategic AI asset. AI agent control and governance is the most strategic element of the above for buyers. But AI orchestration is closest to where work gets done, so more economic activity flows through it, resulting in a larger total addressable market for tech vendors to chase after, especially in the near term. That may tempt Salesforce to chase after orchestration at the expense of Agent Fabric when it should be doubling down on Agent Fabric. This is especially likely so long as the Agentforce and MuleSoft business units remain siloed from each other. Agent Fabric has a strong lead compared to the competition. Will Salesforce expand that lead or blow it? MuleSoft's customers and competitors anxiously wait to find out. If you'd like to discuss Salesforce's strategy more deeply with me, schedule a guidance session or inquiry. See david mooter at: Technology & Innovation Forum East November 4-5, 2026, New York City Cut through CX tech noise and buy smarter. I walked into my first Dreamforce as a skeptic. Salesforce has made several attempts at IT service management, and none of them have provided the technical depth needed by IT teams. After a single briefing on the latest product and then months of silence, I had filed the effort under another short-lived experiment. The pre-announcements [...] As AI systems become more autonomous, security alone isn't enough to build trust. Learn how responsible AI helps organizations govern AI agents, manage emerging risks, and keep systems aligned with business goals and values.