AI is moving fast and most people are falling behind. 

The Rundown AI keeps you ahead of the curve. 

It's a free AI newsletter that keeps you up-to-date on the latest AI news, and teaches you how to apply it in just 5 minutes a day.

Plus, complete the quiz after signing up and they’ll recommend the best AI tools, guides, and courses — tailored to your needs.

Sign up to start learning.

Issue #86

As MCP turns APIs and other capabilities into tools that AI agents can discover and use, a new question is emerging: how do you monetise access to those capabilities? Bill Doerrfeld explores different approaches to MCP monetisation and what they could mean for API providers looking to turn agent access into a viable business model.

As enterprises move from experimenting with individual AI agents to managing multiple agents, tools, models and workflows, a new infrastructure challenge emerges: how do you govern all of them consistently? Tori Jeffcoat explores the case for an agent control plane that can provide centralised visibility, governance and management across an increasingly distributed agent ecosystem.

As AI agents gain access to enterprise systems, knowing who an agent is is no longer enough. Organisations also need to determine what that agent should be allowed to do at the moment it makes a request. Sudeep Goswami breaks the growing complexity of agent authorisation standards down into two fundamental questions: how does identity arrive, and where is the authorisation decision made? The distinction reveals why standards such as ID-JAG and AuthZEN address different parts of the problem.

As AI accelerates development, APIs can start changing faster than teams can test them. SmartBear put its own drift-detection capability to work on a large, mature API, and uncovered mismatches between the OpenAPI definition and the behaviour of the running service. The interesting part is what those mismatches revealed: not just documentation gaps, but genuine bugs and behaviours that existing tests had missed.

OpenRouter has made model access remarkably simple, but enterprise AI introduces a different set of requirements. Solo.io argues that organisations need more than model routing when agents are handling sensitive data, accessing tools and taking actions across enterprise systems. Christian Posta explores what changes when enterprises build their own AI gateway, particularly around identity, authorisation, credential management and controlling agent actions across LLM, MCP, API and A2A traffic.

Local MCP servers give AI agents direct access to developer machines, but that proximity to files, credentials and corporate systems creates a security challenge that is easy to overlook. MCP Manager explores how a gateway can bring visibility, access controls and auditing to local MCP deployments, including controlling which tools can run and what sensitive data they can access. Becky Brooks offers a practical look at an emerging question for enterprise MCP adoption: how do you govern MCP servers that never leave the developer’s machine?

As AI coding agents move from interactive prompting to unattended loops, the challenge is no longer simply how much agents cost to run. Arjun Iyer argues that the real constraint is the quality and speed of the feedback agents receive. Arjun Iyer explores why local tests and CI can leave agents either fast but poorly informed, or accurate but slow, and why production-like validation environments could change that trade-off. It raises an important question for platform teams: if agents are writing and retrying code autonomously, can your verification infrastructure keep up?

An MCP tool call succeeding does not necessarily mean the user got what they needed. Ben Hudson explores a more meaningful question for MCP teams: how do you measure whether an MCP server is actually delivering value?

AI agents are beginning to move beyond analysing compliance information and into actually configuring the systems that enforce it. Sumsub explores how MCP and agent skills could allow compliance teams to turn policies and regulatory requirements into proposed AML/KYC configurations.

As organisations move AI from experimentation into production, governance becomes less about having a policy and more about controlling how AI is developed, deployed and used. Alisa Abramova and Holly Fang’s guide looks at the key elements organisations need to consider when building an AI governance framework. The piece explores the relationship between risk management, compliance, accountability and responsible AI, offering a practical perspective on how organisations can bring structure to increasingly complex AI deployments.

The API Platforms For Scale Newsletter curates intelligence on the trends, technologies, and practices shaping API platforms.