Insights

Writing for people who have to commission the work.

These notes are published by Phigz. They explain how to scope AI products and agentic workflows without invented statistics or a promise the engineering cannot keep.

Delivery

From MVP to production: building AI products that businesses can actually use

An MVP is a first release, not a performance. The core path, the accounts, the failure behaviour and a way to operate it have to be real on day one.

Phigz · 23 September 2026 · 2 min read

AI-first engineering

What is AI-first product engineering?

AI-first does not mean asking a chat tool to generate a website. It means the product and the engineering around it are designed for model behaviour, review and release.

Phigz · 9 September 2026 · 2 min read

Product engineering

How to add AI to an existing SaaS product without rebuilding everything

The useful AI feature lives inside a workflow you already have. It does not require a new platform, and it should not take the product down when the model is unavailable.

Phigz · 26 August 2026 · 3 min read

Agentic systems

AI agent vs chatbot: what should your business actually build?

A chatbot answers. An agent is asked to carry a task through your systems. Most teams need one of those, and they are not interchangeable.

Phigz · 12 August 2026 · 3 min read