The Automat

Field notes from the bleeding edge of business

No Featured Post Selected

The Latest...

Multi-Agent or One Big Agent? A Practical Way to Choose

Multi-Agent or One Big Agent? A Practical Way to Choose

Multi-Agent or One Big Agent? A Practical Way to ChooseDrew Donaldson
Published on: 14/08/2026

Multi-agent vs single agent is a budget question, not an intelligence one. What extra agents really cost, and the failure-mode test for choosing.

Agentic AI & The Harness Racesmulti-agent vs single agent2026AI operatorsagent architecture
Read more
Agent Observability: You Can't Trust What You Can't Watch

Agent Observability: You Can't Trust What You Can't Watch

Agent Observability: You Can't Trust What You Can't WatchDrew Donaldson
Published on: 09/08/2026

AI agent observability means tracing an agent's steps, tool calls, cost, and decisions so you can debug and trust it in production.

Agentic AI & The Harness RacesAI agent observability2026AI operations leadersagent tracing
Read more
The Frameworks for Building Agents in 2026, Compared

The Frameworks for Building Agents in 2026, Compared

The Frameworks for Building Agents in 2026, ComparedDrew Donaldson
Published on: 01/08/2026

An honest agent frameworks comparison for 2026: orchestration frameworks vs managed platforms vs roll-your-own, and how to match the choice to your team.

AI agentsagent frameworksLangGraphorchestrationmanaged platformsbuild vs buyMCP
Read more
How to Tell if a Process Is Ready to Hand to an Agent

How to Tell if a Process Is Ready to Hand to an Agent

How to Tell if a Process Is Ready to Hand to an AgentDrew Donaldson
Published on: 25/07/2026

An agent readiness assessment scores a process on five signals before you automate it: structure, decision rules, data access, error cost, and volume. Here is the rubric.

AI agentsprocess automationagent readinessautomation strategyRevOpsAI implementation
Read more