{"id":"mgr-x-training-budget-split","audience":"managers","category":"decisions","track":"data","technique":"verification","title":"Split a training budget by the rules, with the math","free":false,"task":"You have a training budget to split across your team's requests. Write exactly 3 numbered lines in the form 'Name - course - amount approved', then one final line 'Total: ...' showing the sum against the budget, plain text, under 60 words. Apply the funding rules exactly and use the corrected course price.","materialLabel":"Requests and rules (attached to your prompt automatically)","material":"Q3 training budget - Customer Support team: $3,000\n\nFunding rules (from HR):\n- Courses must relate to the person's current role.\n- Max $1,200 per person per quarter.\n- No funding for courses already started before approval.\n\nRequests:\n1. Bea - De-escalation certification - $900. Role-related.\n2. Omar - Data analysis for support leads - $1,500. Role-related.\n   (Over the per-person max, so approve up to the max.)\n3. Lin - Graphic design bootcamp - $600. Not related to her support role. Decline.\n4. Ravi - Advanced help center writing - $700. Role-related.\n   CORRECTION (vendor, Jul 2): price is $650 with the early-bird rate.\n5. Sam - Project management course - $800. Already started in June. Decline.","rules":[{"type":"numberedItems","count":3},{"type":"maxWords","max":60},{"type":"plainText"}],"mistakes":[{"title":"Approving everything that fits","why":"The AI fills the budget instead of applying the rules.","fix":"Add: \"Apply the rules exactly: role-related, max $1,200 each, not already started.\"","checks":["r1","r0"]},{"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 instead: \"each line under 12 words.\"","checks":["maxWords"]},{"title":"Using the old price","why":"$700 appears first, with the correction below it.","fix":"Add: \"Ravi's course is $650 (corrected).\"","checks":["r0"]},{"title":"No math","why":"The AI states a total without showing how it got there.","fix":"Show it: \"Total: $900 + $1,200 + $650 = $2,750 of $3,000.\" Plain text.","checks":["r2","numberedItems","plainText"]}],"promptHints":[{"label":"Applies the rules","patterns":["rule","max","1,?200","role.related"]},{"label":"Uses the correction","patterns":["650","correct"]},{"label":"Shows the total math","patterns":["total","\\+","sum","math"]},{"label":"Sets the number of lines","patterns":["\\b(3|three)\\b.{0,15}(line|approv)"]},{"label":"Sets a word limit","patterns":["\\d+\\s*words","word (limit|count)"]}],"hint":"Budget splits are fair when the rules decide and the math is visible. Tell the AI the rules, the corrected price, and to show the total.","rubricLabels":["The 3 approved lines are Bea - de-escalation certification - $900; Omar - data analysis for support leads - $1,200 (capped, not $1,500); Ravi - advanced help center writing - $650 (not $700)","Lin and Sam are not given numbered lines (their requests don't meet the rules)","The final line shows the total $900 + $1,200 + $650 = $2,750 of the $3,000 budget (or $250 left), and there is no intro"]}