A content pipeline that publishes itself

An agent-run pipeline that researches, drafts, reviews, and publishes directory content at a scale a small team couldn't touch by hand.

Metrics in this case study are marked [PENDING APPROVAL] until signed off by the owner. Do not publish unverified numbers.

The problem

Agaves.pro is a growing directory in the agave-spirits world. It needed a steady stream of accurate, well-structured content — brand profiles, articles, events — faster than a person could research and write it, without the quality dropping into generic AI filler.

The system we shipped

A production pipeline, not a prototype:

Research agents pull and structure source material.
Drafting agents turn it into on-brand content.
Human review stays in the loop as the approval gate.
Publishing pushes the approved piece into the live CMS.

Every piece carries its sources and a confidence signal, so review is fast and the output is defensible — the opposite of a black box.

The result

[PENDING APPROVAL: ~900+ articles produced through the pipeline.] The point isn’t the count — it’s that one operator plus an agent library runs a content operation that would otherwise need a team.

This is the model SDintelligence sells: execution leverage, shipped to production.