Fuzzy Matching

Identifying the Need for Fuzzy Matching

Understand Common Challenges

Traditional audit processes rely on exact matches, such as invoice numbers. However, bank statements often lack these identifiers, requiring a more flexible approach to link documents.

Determine Supporting Data

Identify other fields that can help triangulate a match, such as:

• Invoice Total
• Vendor Name

[!TIP] Using a combination of fields, like vendor name and invoice total, reduces false matches that can occur when relying on a single non-unique value.

Setting Up Fuzzy Matching

Select the Match Field

Open your match settings and choose the field you want to use for the secondary match, such as Vendor Identification.

Enable Fuzzy Matching

1. Click the three dots on the right side of the match option.
2. Select Variance from the menu.
3. Enable the Fuzzy Matching toggle. This feature deploys AI to identify similar values that are not exact matches.

Providing Context for the AI

Add Custom Instructions

In the fuzzy matching settings, provide instructions to help the AI understand potential variations in the data. For example:

• "Look for abbreviations and other formats of the vendor name, such as adding the word 'Inc.' or 'Corp.'"
• "Consider name reversals (e.g., first name, last name vs. last name, first name)."

[!CAUTION] Clear, descriptive instructions are critical for accurate AI identification, especially when dealing with complex datasets like payroll registers.

Reviewing Results in the Workpaper

Execute the Match

Run the matching process and navigate to your Workpaper.

Verify Results

Check the results in the matching column for your bank statement. Agentive will display "Yes" for matches identified through fuzzy logic, even if the source data contains abbreviations (e.g., "Microsoft Azure" matching to "Microsoft Corp") or variations in the vendor name (e.g., "ProTrain Solutions" matching to "P-Train Solutions").