/ featured · this week
View all dispatches →
An AI agent is software that takes a goal, plans a few steps, and uses other systems to finish the job. That last part is what makes it different from a chatbot, and it is the part that gets...
Aug 31, 2026/ recent · long reads
Browse long reads →
Google's first page for AI agent governance describes controls in exhaustive detail and dates almost none of them. An AI agent, for this purpose, is software that takes actions inside your real...
Aug 30, 2026
An AI agent governance policy template is a document with numbers in it: dollar ceilings, named systems, named tools, one named approver, and a retention period. An AI agent, for the purposes of that...
Aug 30, 2026/ selected notes
All dispatches →01 / AI Agents
Every published AI agent ROI calculator gets the arithmetic right, and the arithmetic is not the trivial part: benefits minus costs over costs, then payback, then net present value for the finance...
Aug 29, 2026
02 / Playbooks
Multi-step agent workflow evaluation is a spending decision disguised as a test suite. The unit under test is never the model alone; it is the harness plus the model, meaning the scaffold that holds...
Aug 24, 2026
03 / AI Agents
An AI agent evaluation plan template is a fixed set of sections that turns scattered agent testing into one signed decision about whether a workflow changes hands. Eight sections carry the whole...
Aug 19, 2026
/ the feed · everything we've published
Showing: Playbooks
Multi-step agent workflow evaluation is a spending decision disguised as a test suite. The unit under test is never the model alone; it is the harness plus the model, meaning the scaffold that holds...
Kurt Fischman · Aug 24, 2026
Nothing about a deployed chatbot upgrades into an agent, which is why the promotion usually arrives as a rebuild disguised as a rename. A chatbot is scoped to a conversation: it matches an incoming...
Kurt Fischman · Aug 18, 2026
> An AI agent implementation playbook is the ordered sequence that moves an agent from a chosen use case to trusted, unsupervised production. Its seven stages (scope, score, design, build, evaluate,...
Kurt Fischman · Jun 22, 2026
> AI agent evaluation is the testing discipline that decides whether an agent can be trusted to own a business workflow. It runs on three surfaces: reliability (does the agent complete the job),...
Kurt Fischman · Jun 10, 2026
> Choosing the right AI agent use case means passing each candidate workflow through three gates in order: value (is the job worth owning), risk (can the business afford the worst day), and...
Kurt Fischman · Jun 10, 2026
> Designing an agentic workflow means specifying five stages in order: intake (the trigger and the inputs), reasoning (the decision and its limits), action (what the agent may write and where),...
Kurt Fischman · Jun 8, 2026
An AI Agent System is the productized operational layer that wraps one or more AI agents, the automation underneath them, the governance around them, and the integrations into a company's real...
Kurt Fischman · May 26, 2026
Reimagine your business with Marshal on the team.