The journal
Lessons from practice and reflections on product, technology and organisation.
Featured
From idea to production: how we build with AI
6 min read
Inside the building of The Playground: how agent context, product decisions, checks and lessons from use connect an initial idea to a product in production.
Read the article6 min read
Agentic workflows: from answers to getting work done
How to delegate a business process to AI agents: tools, permissions, human approval and outcome tracking, illustrated through a customer change request.
Read the article4 min read
With AI, where do we learn to exercise judgement?
When agents produce and review work, how do people learn? Ways to preserve understanding, discussion and opportunities to develop judgement within teams.
Read the article4 min read
AI makes building easier. Finding a market is still the barrier.
Lessons from The Playground: why easier development with AI still leaves distribution, adoption, switching costs and product-market fit to work through.
Read the article4 min read
Organising AI agents’ work without piling on tools
Build an agentic software factory around your product: useful context, clear permissions, complementary checks and responsibilities, without needless complexity.
Read the article4 min read
From demo to everyday use: what still needs checking
User journeys, permissions, data and operations: the checks that help a product built with AI move from a successful demo into reliable everyday use.
Read the article4 min read
Understanding what you build when you no longer write the code
Architecture, rules and checks: how to stay in control of a product when AI agents write the code, drawing on a practical example from The Playground.
Read the article4 min read
AI training at work: start with what teams need to be able to do
Design AI training around the work teams actually do: role-specific skills, realistic exercises, evaluating results and applying what people learn over time.
Read the article4 min read
Experimenting with AI: what decision do you need to make?
How to scope an AI experiment around a decision: choose representative cases, assess quality and cost, and define the evidence needed to decide what comes next.
Read the article4 min read
Your teams use AI. What changes in your organisation?
AI changes what teams can do. Explore how responsibilities, decisions and the product lifecycle can evolve by looking at the work people actually do.
Read the article