Section 4: Excel/Google Sheets, data analysis and reporting

Scenario 32: Cleaning up names, dates and duplicates in a customer list

Level: Beginner · Tools: Excel Copilot in Excel, built-in **@clean-data** skill

Who and in what situation

Murat, a customer service coordinator at a logistics company in Ankara, is consolidating delivery records from different branches. The same company name appears several times with differences in spacing, capitalization, date formats and duplicate records.

How it is done today

Murat opens filters and searches for similar names, removes extra spaces and manually corrects date formats. He then investigates why the PivotTable counts some records separately.

How to do it with AI

  1. He opens the Copilot icon in Excel and types @clean-data at the start of the prompt. This skill checks for spaces, inconsistent text and number formats, inconsistent dates, duplicate values and incorrect data types. [74]
  2. First, he says, “List the findings by category, example cell and potential impact”; he does not request automatic corrections.
  3. He checks every suggested change against the customer code, particularly whether two customers with the same name are actually the same record. The skill explains its suggestions and offers selectable corrections. [74]
  4. He selects the appropriate suggestions and accepts them with Apply, then compares the unique customer codes and total row count with the original copy.

Example prompt

Review {the delivery tracking table} using @clean-data. List extra spaces, differences in company name and city spelling, date formats, duplicate values and incorrect data types. Do not make changes yet; state what needs to be checked for each suggestion.

Evidence and real-world example

No published independent case study was found for this logistics scenario. The basis is the checking and selective application workflow Microsoft provides for its @clean-data skill. [74] ChatGPT’s official documentation also documents checks for missing values, outliers, duplicates and incorrect data types. [75]

Points to consider

Microsoft states that the @clean-data skill is still being rolled out to production and may not be available to everyone. [74] The same name does not automatically mean the same customer; base merging decisions on the approved customer code and recount totals after cleaning.

Sources for this scenario

  1. Use built-in skills with Copilot in Excel (support.microsoft.com)
  2. Extracting insights with ChatGPT data analysis (help.openai.com)