Same meeting-notes assistant, redesigned after the first budget review. What changed? Before: one premium lane for every meeting. After: modality and risk decide the route. Modality moved first: speech work goes to an audio or transcription model family because a text reasoning model is not automatically better at hearing the input. Routine summaries got a cheaper default: a fast workhorse handles ordinary status calls after local examples show acceptable factual coverage. Ambiguity still gets capability: decision conflicts and unclear owners escalate because task-model fit protects the cases where mistakes change outcomes. Right-sizing is not downgrading; it is routing each sub-task…
Read the full lesson
Sign up free: one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in