AI + Product

Where product leadership meets AI building

I don't just advise on AI strategy. I build with AI tools every day — production data pipelines, multi-agent systems, workflow automation, and purpose-built applications. I bring that practitioner depth to every client engagement.

The Story

Act One: 2015–2024

A decade of product leadership

I founded theProductPath to work with B2B software companies at every stage — from seed-stage startups building their first product team to growth-stage companies scaling their product organization. I wrote 50+ articles documenting real product decisions, led a PM community for 8+ years, and helped dozens of companies ship great products.

Act Two: 2025–Present

Building with AI, not just advising on it

Starting in late 2025, I went deep into AI — not as a reader or evaluator, but as a builder. I shipped production data pipelines for clients, built multi-agent travel planners, created mobile apps, designed personal AI infrastructure, and developed frameworks for teaching non-technical people to build with AI. I now work across 9+ AI tools daily.

What I've Built

Real projects, not slide decks

A selection of AI projects from my practice — ranging from client deliverables to personal infrastructure to published frameworks.

Client Work

Production Data Pipeline

Full ETL pipeline for a client's ERP migration: fuzzy matching, phonetic algorithms, deduplication, and a human-in-the-loop web app for review. Built in ~1 week with AI coding assistants.

Architecture

Multi-Agent Systems

Designed and built a coordinator-dispatcher system with 8 specialist AI agents. Hit context window limits at scale, diagnosed the root cause, redesigned to sequential dispatch, and documented the lesson.

Framework

Burner Apps

Purpose-built, disposable software tools created with AI pair programming. A framework for solving bounded business problems without leaving behind maintenance debt.

Read the article →

Personal Infrastructure

AI Operations Layer

Built a personal AI infrastructure where specialized agents process inbound notifications, extract events, and act autonomously — including a morning briefing delivered before I wake up.

Methodology

Parallel AI Collaboration

A lightweight system for two people to run simultaneous AI sessions on a shared project — using a shared folder, a task file, and a ping system. Packaged as a reusable template.

Read the article →

Teaching

AI Building for Non-Technical People

Designed a structured hackathon playbook for teaching non-developers to build with AI: brainstorming interviews, rescue prompts, coach mode, and scoped project lanes.

AI Thinking

What I'm writing about

Recent posts exploring the opportunities and limitations of working with AI.

AI Blindspot Series

Why Results Matter More Than Actions

AI models don't know if their advice works. The gap between tracking what you asked for and tracking what actually happened is the real frontier.

AI Blindspot Series

The Missing Feedback Loop

AI has no idea whether you used the information it gave you. Until we solve this, every session starts from scratch.

My AI Toolkit

9 tools in active daily use

Claude Code

Codex

Claude Cowork

ChatGPT

Cursor

Claude for Chrome

Notion AI

OpenClaw

Manus

Want to put AI to work in your product organization?

I can help you figure out what's real, what's useful, and what to build first.