The agentic era will not be won by the teams with the cleverest prompt. It will be won by organizations whose systems expose trustworthy, discoverable capabilities that agents can safely use.

That is the premise of The Agentic Architect: architecture patterns for agentic systems people can trust.

Models matter. But a capable model without the right context, tools, identity, and controls is still a limited agent. The hard—and interesting—work is building the architecture around intelligence so that it can understand an objective, act responsibly, and earn trust.

Who this is for

This is for architects, API and platform teams, product leaders, and builders who are moving from AI demos to durable systems. It is a place to connect the technical foundations of the enterprise with the emerging ways agents can use them.

What we’ll explore

  • Agent architecture: orchestrators, planners, evaluators, and the choices that shape behavior.
  • Agent-accessible capabilities: APIs, events, MCP tools, and context that agents can discover and compose.
  • Trust and operations: identity, authorization, governance, observability, reliability, and cost control.
  • Field notes: practical lessons from designing and operating agentic systems in the real world.

Start here

The first substantive essay, From Pong to Minecraft: How GenAI Is Redefining the User Experience, lays out the central shift: the next generation of software is built around user intent and composable capabilities, not just predefined workflows.

This is a living set of notes. The goal is not to chase every new model release, but to understand the architectural patterns that let intelligent systems become genuinely useful.

Frequently asked questions

What is The Agentic Architect about?

It’s a blog about the architecture underneath useful, trustworthy AI agents—APIs, events, MCP, context, memory, identity, and governance—rather than model selection or prompt engineering. The premise is that a capable model without the right context, tools, and controls is still a limited agent.

Who is The Agentic Architect written for?

Architects, API and platform teams, product leaders, and builders who are moving from AI demos to durable, production systems, and who want to connect the technical foundations of the enterprise to the emerging ways agents can use them.

Where should I start reading?

Start with From Pong to Minecraft: How GenAI Is Redefining the User Experience, which lays out the central shift this site explores, or jump to Start Here for a guided reading order through the six essays on agent architecture.