Blog
Field notes on AI product planning, prompt engineering for PMs, and the pipeline behind VibeMap.

product-strategy
A lightweight business case template for builders. Pressure-test any product idea in 30 minutes, with a worked example, kill criteria, and the self-deceptions to avoid.

database-design
Turn user stories into a production-ready PostgreSQL schema in five steps: extract nouns, define entities, map cardinality, add constraints. Full worked example with SQL, plus a free database schema generator.

acceptance-criteria
Learn Gherkin acceptance criteria in plain English: 12 complete Given/When/Then examples across auth, payments, search, and edge cases, plus a bad-criteria rewrite gallery and a checklist for handing scenarios to AI coding agents.

user-stories
Learn how to write user stories with the INVEST criteria: 8 real examples, bad-to-good rewrites, slicing tactics, and a checklist for AI-assisted builds.

acceptance criteria
LLMs write acceptance criteria that cover the happy path and quietly omit everything else. Here are the 12 edge-case categories they miss by default, the prompt that forces coverage, and how to review generated criteria.

AI app architecture
How to produce a scalable Next.js or React project structure, relational schema, and component inventory from a prompt without getting generic scaffolding. With working examples and prompt templates.

AI code generation
A catalogue of real failure modes in AI-generated code: component duplication, schema drift, missing failure paths, and more, with code examples and what actually prevents each one.

AI PRD generator
Three free copy-paste PRD templates (lean, standard, and agency-grade), the prompts that produce each from a plain-English idea, and an honest look at where the PRD format breaks down.

AI product planning
The exact pipeline for turning a plain-English product idea into a complete app specification: personas, features, user stories, acceptance criteria, schema, and pages. With prompt templates and working examples.

persona generator
How to generate credible user personas with AI without the 'Sarah the Marketer' stock characters. Includes the exact prompts, a five-question audit, and an honest take on when AI personas are not enough.

prompt engineering
Ready-to-use prompt templates for the work PMs actually do with LLMs: customer segments, prioritisation, competitive research, interview synthesis, stakeholder updates, and retros. With the constraints that stop output drifting generic.

AI product management
Getting user stories out of an LLM is easy. Getting ones you would hand to a developer takes a second pass. Here is the generate, critique, and refine loop, plus the failure signatures specific to AI-written stories.

AI product management
Everyone quotes the same statistic about planning saving money. It is shakier than it looks. Here is what the research on AI-assisted development actually supports, and what it does not.

vibe coding
Vibe coding was coined to describe throwaway weekend projects. It got adopted by people shipping production SaaS. Here is the difference, why it matters, and how to tell which one you are doing.