{"id":"fpa-driver-forecast-assumptions-in-words","audience":"fpa","category":"finance-work","track":"finance","level":"intermediate","technique":"audience","title":"Write forecast drivers as plain-English assumptions","free":false,"task":"The leadership team wants the FY27 subscription forecast drivers in plain words, not spreadsheet logic. Write exactly 5 numbered assumptions, one per driver, each a single sentence with the value and its source in parentheses at the end. Use the latest value for each driver. No formulas, \"=\" signs or cell references. Plain text, under 120 words.","materialLabel":"Source material (attached to your prompt automatically)","material":"FY27 SUB REVENUE DRIVERS (tab: Drivers)\n\nC4 Starting customers (Jan 1): 5,000 - source: billing system, Oct 31 count\nC5 New customers per month: 150 - source: Sales Ops plan\nC6 Monthly churn: 1.2% - source: trailing 12-month average, Finance\nC7 Average monthly price: $80 through March, $84 from April 1 - source: Pricing committee\nC8 Monthly expansion from upgrades: 0.5% - source: Customer Success\n\nLogic: C11 = C10*(1-C6)+C5; revenue = C11*C7*(1+C8)\n\nUpdate, Nov 1 (Sales Ops): after the SDR team was cut from 10 to 8, new customers per month is revised to 120. Plan refresh pending.\n\nComment on C6: \"Churn spiked to 2% in July because of the outage, but that's already in the trailing average.\"\n\nVersion: v3, saved by D. Ruiz.","rules":[{"type":"numberedItems","count":5},{"type":"plainText"},{"type":"maxWords","max":120}],"mistakes":[{"title":"Using the stale driver","why":"The AI copies 150 from the driver table and misses the Nov 1 revision.","fix":"Add: \"Use the latest value for each driver; new customers is 120 after the Nov 1 update.\"","checks":["r0"]},{"title":"Dropping the sources","why":"The AI states assumptions without saying who owns them, so nobody knows whom to ask.","fix":"Say: \"End each line with its source in parentheses.\"","checks":["r1"]},{"title":"Echoing the spreadsheet","why":"The AI repeats cell references and formulas that leaders can't read.","fix":"Say: \"No formulas, = signs or cell references.\"","checks":["r2"]},{"title":"Trusting the AI to count words","why":"AI is bad at counting and runs long, adding lists and formatting you didn't ask for.","fix":"Cap each part (\"each line under 15 words\") and state the format: \"plain text, no bold.\"","checks":["numberedItems","plainText","maxWords"]}],"promptHints":[{"label":"Uses the latest values","patterns":["latest","revis","120","nov 1"]},{"label":"Cites sources","patterns":["source","parenthes","owner"]},{"label":"Bans spreadsheet logic","patterns":["no formula","cell ref","no =","plain (words|english)"]},{"label":"Fixes the count","patterns":["\\b(5|five)\\b.{0,20}(assumption|line|numbered)"]},{"label":"Sets a length limit","patterns":["\\d+\\s*words","word (limit|count)","under \\d+"]}],"hint":"Leaders need words, not cells, and the latest numbers. Tell the AI where each line's source goes.","rubricLabels":["New customers per month is 120 (Sales Ops revision of Nov 1), not 150","Each of the 5 assumptions ends with its source in parentheses (billing system, Sales Ops, Finance, Pricing committee, Customer Success)","Contains no formulas, \"=\" signs or cell references like C5 or C11"]}