{"id":"exec-operating-review-variance-flags","audience":"executives","category":"leadership-work","track":"leadership","level":"intermediate","technique":"verification","title":"Flag the variances in an operating review that need a decision","free":false,"task":"Review the monthly figures. For each line, compute the variance (actual minus budget, and percent of budget, rounded to a whole number) and flag it \"REVIEW\" only if the absolute variance is over 10% of budget, otherwise \"ok\". Write exactly 5 numbered lines like: 1. Line - actual vs budget - variance $ and % - REVIEW/ok. Then a final line: Biggest: the line with the largest variance in dollars, with no cause guessed. Show the math.","materialLabel":"October figures (attached to your prompt automatically)","material":"October operating review (all in $000)\n\nLine | Budget | Actual\nRevenue | 1,200 | 1,104\nCost of goods | 480 | 504\nSalaries | 360 | 372\nMarketing | 80 | 62\nSoftware | 50 | 59\n\nNotes: Marketing underspend is a timing issue per the VP, but I haven't confirmed it. Revenue figure was revised from 1,090 to 1,104 on Nov 3 after one late invoice. Software overrun came from a one-time license fee.\n\nRule: percent = variance / budget. For costs a positive variance is an overspend.","rules":[{"type":"numberedItems","count":5},{"type":"maxWords","max":170}],"mistakes":[{"title":"Using the stale revenue","why":"The AI uses 1,090.","fix":"Say: \"Revenue was revised to 1,104.\"","checks":["r1"]},{"title":"Flagging by feel","why":"The AI flags the largest dollar items.","fix":"Say: \"REVIEW only if over 10% of budget.\"","checks":["r0"]},{"title":"Guessing causes","why":"The AI says marketing is timing.","fix":"Say: \"Don't guess causes; the VP's timing claim is unconfirmed.\"","checks":["r2"]},{"title":"Skipping the math","why":"The AI states variances without working.","fix":"Say: \"Show actual vs budget and the variance.\"","checks":["r0","numberedItems"]},{"title":"Trusting the AI to count words","why":"AI is bad at counting. Ask for a total word limit and it often lands over it.","fix":"Cap each line: \"each line under 25 words.\"","checks":["maxWords"]},{"title":"Trusting the AI to count words","why":"AI is bad at counting. Ask for a total word limit and it often lands over it.","fix":"Cap each part instead, for example: \"each line under 12 words.\"","checks":["maxWords"]}],"promptHints":[{"label":"Line format","patterns":["1\\.\\s*line","actual vs budget"]},{"label":"Threshold","patterns":["10%","over 10"]},{"label":"Percent rule","patterns":["variance / budget","percent"]},{"label":"Latest data","patterns":["revised","1,104"]},{"label":"No causes","patterns":["don'?t guess","no cause"]}],"hint":"Show the line format, the 10% rule and the formula, and forbid guessed causes.","rubricLabels":["Each of the five lines shows actual vs budget, the dollar variance and percent, and flags consistent with the rule: Revenue (-96, -8%) ok, Cost of goods (+24, +5%) ok, Salaries (+12, +3%) ok, Marketing (-18, -23%) REVIEW, Software (+9, +18%) REVIEW","Revenue uses the revised figure of 1,104 (not 1,090)","The final line starts with \"Biggest:\" and names Revenue (96) as the largest dollar variance with no guessed cause"]}