In the era of AI agents and automated integration, the roles of platform engineers and API managers are rapidly converging. As enterprises seek to break down technological silos while maintaining governance, a new approach to API ecosystem management is emerging—one that balances centralized control with the flexibility to support diverse technology stacks.
At Boomi World 2025, I spoke with Markus Mueller, Boomi’s Global Field CTO for API Management, about how the rise of AI agents fundamentally changes API design, governance and management requirements. His insights and Boomi’s recent announcements around Model Context Protocol (MCP) support reveal important principles for platform engineers looking to build more resilient and flexible API ecosystems.
The AI Agent Imperative for API Governance
“If you have done proper API management in the past, you are good to go,” Mueller explained when discussing how Boomi’s automatic exposure of APIs as MCP server endpoints will affect organizations. However, he quickly acknowledged the reality: Most enterprises haven’t implemented comprehensive API governance, instead relying on human relationships to patch inconsistencies.
“With humans, if they don’t understand something, they just call you,” Mueller noted. “They pick up the phone, call you, say, ‘I don’t understand the documentation.’ An agent cannot do that. The agent will try to figure out what it should do and then do it.”
This represents a fundamental shift for platform engineers. While human developers might work around poorly documented or inconsistently designed APIs through direct communication, AI agents lack this capability. The result is what Mueller describes as “the pressure for customers to get their APIs right and to get governance on APIs right, because they don’t want to deal with the fallout of these misinterpretations or misunderstandings.”
The Domain Translation Challenge
One particularly illuminating example Mueller shared involves the ambiguity of business terminology across domains:
“The example I always use is an ‘account’—if you’re an IT guy, ‘account’ is a user account of some system. If you come from sales, that’s your customer,” he explained. “Now, you tell the agent to create an account, which will look for tools. And it says, ‘Oh, which tools can create an account?’ And there will be two, and it has to choose now.”
This highlights a critical challenge for platform engineers: Domain-specific terminology must be explicit and unambiguous when AI agents are involved in API design. As Mueller succinctly put it: “Descriptions like ‘This is the account’ will not cut it.”
Federation Without Fragmentation
Boomi’s approach to API management demonstrates a key platform engineering principle: Ecosystem federation without fragmentation. The company’s recent announcements reveal how this works in practice.
The introduction of native support for MCP throughout Boomi’s architecture is particularly significant. By automatically exposing APIs as MCP server endpoints via Boomi API Management, the platform enables AI agents to discover and interact with enterprise systems while maintaining governance controls.
Moreover, Boomi’s introduction of a new MCP gateway for tool aggregation and discovery demonstrates how platform engineers can build federated systems that maintain centralized governance while avoiding vendor lock-in. This gateway approach enables enterprises to combine diverse tools and technologies under a unified governance framework.
During his keynote address, Boomi CEO Steve Lucas described APIs as “the lingua franca” of the agent world, underscoring their fundamental importance in creating interoperable systems. For platform engineers, API design and governance are no longer secondary concerns but central to enabling AI-driven automation.
The End of “Getting Away With It”
Perhaps the most significant insight for platform engineers is Mueller’s observation about accountability: “You can get away with it by human relationship. I think that’s what changed here, because you can’t have a relationship with your agent.”
He elaborated with a practical example: “Consider the case where you may have few consumers. Then you, as a publisher, you know them, and you can just give them a ring and say, ‘We’re thinking about changing this and this, are you okay with it? Can we do it?’ You don’t have that with the agents.”
This represents a fundamental shift in platform engineering responsibility. The informal communication channels and relationship-based exception handling that have long characterized enterprise IT must give way to more rigorous, systematized API design and governance approaches.
“Agentic agents necessitate a tighter alignment between platform engineering and API management strategies. A well-engineered platform is essential for a discoverable, secure and consistently accessible API catalog that agents can effectively utilize,” said Mitch Ashley, VP and practice lead, DevOps and application development at The Futurum Group. “By prioritizing API design principles that cater to the needs of AI, organizations can fully leverage the power of AI agents while minimizing integration efforts.”
Practical Approaches for Platform Engineers
Based on the insights from Mueller and Boomi’s broader strategy, platform engineers should consider several approaches to create flexible API ecosystems that support diverse technology stacks:
- Implement domain-specific vocabularies: Create clear, unambiguous API descriptions for terminology differences across business domains
- Adopt federation architectures: Use technologies like Boomi’s MCP gateway to aggregate and govern diverse tools while maintaining centralized visibility.
- Design for agent consumption: Assume APIs will be consumed by AI agents that cannot ask clarifying questions or understand implicit context.
- Standardize documentation: Implement consistent documentation practices that include examples, edge cases and clear parameter descriptions.
- Implement API lifecycle management: Establish formal processes for versioning, deprecating and communicating API changes.
The Future of Platform Engineering and API Management
As AI agents become more prevalent in enterprise architectures, the distinction between platform engineering and API management will continue to blur. Platform engineers will increasingly need to consider how their infrastructure choices affect API consumption patterns, while API managers will need to understand the platform implications of their design decisions.
Boomi’s strategy of combining integration, API management and AI agent orchestration under a unified governance framework offers a blueprint for this convergent future. By adopting similar principles, platform engineers can create ecosystems that enable innovation while maintaining the governance controls necessary for enterprise operations.
The convergence of these disciplines ultimately serves the broader goal of breaking down enterprise silos — not just between applications and data sources but also between the teams and governance structures that have traditionally managed them separately.