Introduction

Chartered Accountants often play a critical role in forensic investigations. As part of these investigations, email conversations and system files of the suspect are collected and reviewed. While there are several well-established ESI (Electronic Stored Information) tools available in the market, they are often expensive and not budget-friendly for many organizations.

Among all file formats, Excel review plays a particularly crucial role. Consider a case of kickbacks taken by an employee in exchange for vendor favoritism:

· The suspect may maintain a personal record of kickbacks in an Excel file, hidden under a misleading file name or within a hidden sheet.

· Sometimes the investigation may focus on locating quotes, purchase orders (POs), or goods receipt notes (GRNs), which are also frequently stored in Excel.

Thus, Excel files are often goldmines of evidence—but only if reviewed thoroughly.

Challenges in Reviewing Excel Files

· Sheer Volume Every employee in today’s workplace uses Excel regularly. Systems can easily contain thousands of Excel files, making manual review impractical.

· Hidden Data Suspects often hide data using hidden rows, columns, or sheets. Opening each file and unhiding data manually is not feasible.

· Metadata Importance Investigators may not only need the cell data, but also who created or last modified the file along with timestamps. This metadata often provides crucial leads.

· Limited Level 2 Review Since investigations are time-sensitive and the volume of files is huge, often only sample checks are done by Level 2 reviewers. A large reliance is placed on Level 1 reviewers, increasing the risk of missing critical evidence.

Using AI for Smarter Investigations

With advancements in AI, it is now possible to build custom investigation tools using Python, aided by AI assistants like ChatGPT, Copilot, or Gemini.

I developed a tool that focuses on Excel review to make investigations faster, more comprehensive, and cost-effective.

Key Features of the Excel Review Tool

1. Automated Unhiding of Data

· The tool automatically scans Excel files and unhides all hidden rows, columns, and sheets, ensuring nothing is missed.

2. Comprehensive Search Scope

Once the user:

1. Inputs keywords into a keyword template sheet, and

2. Places all Excel files into the input folder,

the tool scans across multiple layers:

· Cell values – including hidden sheets and rows.

· File names – filenames are often disguised and need scanning.

· Sheet names – can reveal intent (e.g., “Bribes,” “Cash Flow”).

· Metadata – Author and Last Modified By details are extracted.

At the end of the run:

· Files with hits are moved into a Hits folder.

· Files with no hits are moved into a No Hits folder. This segregation saves time for the investigator.

3. Advanced Query Logic

Investigators can use complex search rules similar to a search engine:

· AND → Both terms must appear.

o Example: Audit AND Expense.

· OR → At least one term must appear.

o Example: Gift OR Voucher.

· Nested Rules with Parentheses

o Example: Deepak AND ("Gift" OR "Voucher").

This allows for context-sensitive search strategies, rather than plain keyword matching.

4. Investigator-Friendly Reports

The tool generates structured Excel reports for the reviewer:

· Comprehensive Report (Sheet 1)

o Contains all hits across cells, file names, sheet names, and metadata.

o Columns include:

§ File Path

§ Sheet Name

§ Cell Address

§ Keyword Rule Triggered

§ Hit Type (Cell / Filename / Sheet / Author / Modifier)

§ Author

§ Last Modified By

· Metadata-Only Report (Sheet 2)

o Isolates files where hits were only in metadata.

o Useful for identifying cases where authorship or modification itself raises questions.

o These files are also copied into a MetadataHits folder for quick access.

5. Parallel Processing for Speed

· Files are divided into batches.

· Up to five worker threads process files simultaneously.

· This allows large volumes to be processed in a fraction of the time compared to traditional sequential methods.

Future Enhancements

While the tool is currently built for Excel review, future development can extend its scope:

· Word documents and PDFs – searching for embedded text.

· Image files – using OCR (Optical Character Recognition) to read scanned documents.

· Email files (PST/MSG) – to integrate email review with Excel search.

Conclusion

Excel files often hide the most crucial evidence in forensic investigations. Manual review is not practical given the volume and complexity. This AI-enabled Excel review tool ensures investigations are faster, smarter, and more reliable. By combining comprehensive search, metadata analysis, and speed, it provides Chartered Accountants and investigators with a cost-effective alternative to expensive ESI platforms.