Tasks, Skills, and Tools: The Holy Trinity of AI Agents
Every enterprise AI conversation eventually arrives at the same question: how do we actually build agents that work?
Not …
Read article →// Enterprise AI Engineering
Build enterprise AI that actually works.
Protocols. APIs. Agents. The craft of building production-ready systems for the agentic era.
Enterprise AI doesn't fail because of models.
It fails because of missing contracts, unclear governance, APIs that were never designed for agents, and orchestration patterns borrowed from tutorials instead of production systems.
The Protocol Practice is where enterprise architects and API practitioners learn to build the foundations that make AI work at scale. Deep dives into protocols, best practices for API design, and practical patterns for agent building and orchestration—no hype, just rigor.
Topics
What we cover
Agent Building
From single-agent patterns to multi-agent orchestration. Architectural decisions for agents that collaborate, not collide.
View articles →API Design
Building APIs that are discoverable, self-describing, and agent-consumable—without sacrificing governance and security.
View articles →Governance
API governance, agent oversight, and operational readiness. Speed without guardrails is just technical debt with a longer fuse.
View articles →Protocols
MCP, OpenAPI, AsyncAPI, and the standards shaping agent-to-system communication. What works, what doesn't, and how to implement them.
View articles →Latest Article
The False Economy of Speed: Why Skipping API Governance Leads to Agentic Technical Debt
When every team can deploy an agent, governance becomes the difference between compounding value and compounding chaos. Here’s what enterprises get wrong—and how to fix it.
Read article →// agent-config.yaml
skills:
- inventory.check
- orders.create
- shipping.estimate
governance:
rate_limit: 100/min
audit: enabled
approval: required
skills:
- inventory.check
- orders.create
- shipping.estimate
governance:
rate_limit: 100/min
audit: enabled
approval: required
The best AI systems will be built on the best foundations.
APIs designed with intention. Protocols implemented with rigor. Governance as an enabler, not an obstacle. The Protocol Practice is for practitioners who understand that enterprise AI is an engineering discipline, not a magic trick.
Recent
Latest articles
The False Economy of Speed: Why Skipping API Governance Leads to Agentic Technical Debt
When every team can deploy an agent, governance becomes the difference between compounding value and compounding chaos. …
Read article →MCP Deep Dive: What the Model Context Protocol Actually Does (And Doesn't Do)
Model Context Protocol is gaining traction as the standard for agent-to-tool communication. Here’s a …
Read article →About
From the field, not the sidelines
Written by Markus Müller, Global Field CTO for API Management at Boomi. With deep experience in enterprise integration, API strategy, and the emerging landscape of agentic AI, Markus brings a practitioner's perspective to the protocols and patterns shaping how enterprises build for AI.
This isn't analyst commentary from the sidelines. It's field notes from someone building, advising, and shipping in the enterprise AI space.
More about The Protocol Practice →