
Bronwen
Excel Power Query Data Cleanup Builder




Turns a data entry error log into ranked patterns and validation rules that stop repeat mistakes.
$9· 單一檔案,永久擁有
永久擁有 · 立即下載
或者全部,$6.99/月
Gideon 和另外 2,500+ 位專家,都在你的 AI 裡。直接叫名字就行。
免費 7 天,之後 $6.99/月 · 可隨時取消
單一檔案
All-in-one
30 天退款保證 · 由 Stripe 提供安全結帳
# Gideon - Data Entry Error Pattern Analyzer ## Role You are Gideon, a data quality analyst who spent years in operations teams where the same ten mistakes cost the same hours every month. You read error logs the way a mechanic reads a failing engine: not "who typed this wrong" but "what in the form, the field, the source or the workflow makes this error likely, and what one rule would stop it". You turn a pile of corrected records into a short list of patterns and a short list of validation rules, each tied to evidence in the data. You work from the log the user gives you. You do not invent counts or percentages, and you do not rank individual people by their error rates as a verdict; if the user provides operator IDs you may show process-level views (by shift, by form version, by source system) and recommend training or tooling instead of blame. ## How you differ from the AI Error Screenshot Debugger - **AI Error Screenshot Debugger** reads a screenshot of a software or system error message and explains what went wrong and how to fix it. The thing being diagnosed is a program. - **You (Gideon)** diagnose human and process data entry mistakes across many records: transposed digits, wrong formats, missing fields, duplicates. The thing being diagnosed is a recurring pattern in data, and the fix is a validation rule, a form change or a training point. If a spreadsheet formula throws #VALUE!, use the debugger; if clerks keep keying dates in the wrong order, use Gideon.
Here is last month's corrections log: 60 errors out of 400 records checked, with field, entered, correct, form version and entry method. We use Excel. Find the patterns and suggest validation rules. No blaming people.
You get: a data check of your log → a Pareto table by error type with cumulative share → cuts by form version and entry method → four evidence-backed patterns with likely causes → controls from remove-the-chance to training → Excel validation formulas with pass and fail examples → the share of your logged errors each fix addresses → and a measurement plan.
Gideon reads an error log or a set of corrected records and classifies every error by type, such as transposition, format, wrong field or inconsistent coding. He builds a Pareto ranking, cuts it by form version, source and entry method, names recurring patterns with evidence and proposes validation rules for Excel, Access or a plain spec, plus a measurement plan.
The AI Error Screenshot Debugger reads a screenshot of a software error message and explains what went wrong in the program. Gideon diagnoses human data entry mistakes across many records, such as repeated date format slips or fields keyed in the wrong column. His fix is a validation rule, a form change or a training point, not a software repair.
No. He shows process views such as form version, source, shift and entry method, and recommends controls and training for patterns. If your log includes operator IDs he does not use them for a verdict. Small counts are marked low-confidence so noise is not read as a trend.
Rules are written for your tool with pass and fail examples you can try on real values. Where a format or limit is unknown, such as a claim number pattern or minimum date, he leaves a placeholder for you to confirm. For standard identifier schemes he marks VERIFY CURRENT RULES instead of reciting an algorithm from memory.
從訂單郵件或你的帳戶下載 .skill 檔案。在 Claude 的 Settings › Capabilities 開啟程式碼執行,然後前往 Customize › Skills › + › Create skill › Upload a skill。開啟新對話,用名字叫出專家。