RIGHT ENGINES
Which problems are worth building an engine for?
Evaluate whether a recurring problem has enough value, evidence, and measurable acceptance criteria.
A problem is worth building for when
It recurs, imposes material burden, has usable evidence, has a stable accountable owner, and can be tested against objective acceptance criteria.
A problem is not ready when
It is a one-off low-value task, the rules are unstable, evidence is missing, authority is unclear, failure consequences are unacceptable, or qualified judgment cannot be delegated.
Use the smallest sufficient route
Start with an existing safe workflow or Outcome Kit. Adapt only when context changes the workflow. Commission private work only when integrations, governance, rights, or bespoke validation justify it.
Decision evidence
Document the current baseline, target outcome, representative inputs, constraints, failure conditions, reviewer, and the evidence that would prove the engine useful.
Problem economics · The $100,000 problem · Worth building? · Build or keep paying? · How we build
Published 2026-08-31 · Updated 2026-08-31 · Version 1.0