From Marketing & BI to Building AI Tools That Augment Human Judgment

Hey everyone! 👋

I have a background in marketing, business intelligence, and technical writing. Lately, I've been building open-source tools that turn domain knowledge into explicit, repeatable guidance for AI assistants.

The philosophy behind them is pretty simple: help people make better decisions, don't just automate their work.

A lot of this started during my time as a BI engineer. I was trying to help marketing teams become more data-driven, but neither training sessions nor dashboard walkthroughs seemed to make much difference. What eventually worked was creating documentation that explained not just what the data showed, but why it mattered and how to use it. Combined with AI, that gave people a way to explore the data, ask questions, and build their understanding independently.

That experience got me interested in a broader question: what if we could make our accumulated professional knowledge available to AI in a structured way, without giving up our own judgment?

I've since built three projects around that idea:

• swift-pbi: Helps business teams prototype Power BI data models while giving data professionals tools to validate and maintain them.

• Devil's Analyst: Encodes marketing analytics expertise into checks that challenge questionable interpretations before they become business decisions.

• Second Read: An editorial skill for people who still want to write and revise their own drafts, with a cloud workspace that can adapt to their editorial preferences over time.

All three grew out of things I encountered in my own work, combined with what I've learned through academic studies, professional experience, and a lot of independent reading.

I'm particularly interested in the space between AI automation and human expertise: how we can make AI more useful, explainable, and adaptable without losing the judgment that makes our work meaningful.

I'd love to connect with other people exploring similar ideas, especially around AI agents, knowledge management, writing, and open-source tools.

Curious to hear: are you building AI tools to automate tasks, or to help people become better at doing them?