Writing

I write about building and shipping AI applications from three main perspectives, also referred to as Tracks: engineering, business, and product.

Search Guide

Learn how posts are organised — and how to navigate them.

Track

Engineering

Engineering Deep-Dives Architecture decisions, implementation challenges, and solutions.

Technical Writing General agentic AI insights, industry trends, and emerging patterns.

Business

Business & Operations Practical insights on pricing, distribution, and running a portfolio.

Metrics & Progress Real numbers and honest updates on what's working and what's not.

Product

Product News Build logs, learnings, and updates on shipped AI products.

Community Feedback Iterating on products based on user requests and collaboration.

Series

Blog posts can also belong to a Series, which collects related posts exploring a topic deeply—whether that's a product's journey, a technical subject, or a business theme.

Use the Series filter to browse all posts in an arc together, in the order they were written.

Phase

Product Series posts are organised into Phases of a product's development journey: Strategy, Design, Engineering, Deployment, Maintenance, and Community (feedback).

The Phase filter appears automatically when a product series is selected.

Essential Reading

The 3 most important posts to start with, if you haven't already.

What are AI agents and agentic systems, anyway?

A plain-English explanation of AI agents, agentic systems, and what makes them different from chatbots—plus the production lessons people skip.

Building AI Products Solo — And Why It Works

Why solo doesn't mean small anymore. What AI augmentation changes about speed, quality, and the customer relationship—especially for MSMEs.

Why I build AI systems with code, not no-code tools

No-code AI tools are great for prototypes. Here's why I switched to code-first for production and what that stack actually looks like.

Showing 24 posts

Aug 13, 2026 · 6 min read

GovCon Leads Radar — Building a radar, not another lead list

Why GovCon Leads Radar went API-first and horizontal from day one, instead of chasing a narrow niche.

Product

GovCon Leads Radar Strategy

Aug 13, 2026 · 16 min read

Social Engagement Radar — Sixteen decisions locked before Phase 2 engineering began

How Social Engagement Radar's strategy phase turned a validated vision into sixteen locked decisions across workflow, capability, stack, and risk.

Product

Social Engagement Radar Strategy

Aug 2, 2026 · 7 min read

Your Latency Budget May Be Lying to You

Why cycle-bound products need their own latency model instead of borrowing real-time P95 conventions.

Engineering

Aug 2, 2026 · 11 min read

PolicyForge — Inside the MVP Technical Stack

Inside the week I locked PolicyForge's technical foundation, and the tradeoffs I chose along the way.

Product

PolicyForge Strategy

Aug 1, 2026 · 7 min read

Running a boundary-table test on a full, AI-enabled workflow

The five-question boundary test worked on single AI capabilities. Here's what happened when I ran it across a connected workflow instead.

Product

Aug 1, 2026 · 7 min read

PolicyForge — What mapping our first workflow actually revealed

Mapping PolicyForge's first workflow turned an abstract bet on AI-human boundaries into four concrete decisions.

Product

PolicyForge Strategy

Jul 31, 2026 · 8 min read

PolicyForge — Why the first workflow we build determines everything after

Choosing the one workflow to build first in PolicyForge, and why dependency-free beats impressive.

Product

PolicyForge Strategy

Jul 30, 2026 · 8 min read

OpsPilot — Agent, Copilot, or Both? Locking The Product Category and Autonomy Model

How we settled the Agent vs. Copilot debate for an ops-automation product and what it changed about the build.

Product

OpsPilot Strategy