Checkmate

Inspiration

Surprisingly, the aviation industry then draws many parallels to the current accounting landscape. Firstly, most accounting firms have adopted AI and Machine Learning to flag suspicious transactions for vouching. It is a crucial step in the auditing process to reduce the workload and aid sampling. However, cases like ENRON, 1mdb and closer to home, the Aljunied Town Council, have shown its ineffectiveness. All companies need to do is to discover the parameters of the software and work their way around them. As a result, regulators and researchers have pointed to poor audit quality and insufficient sampling as factors that lead to the occurrence of fraud. As such, we asked ourselves, how can these 2 issues be resolved?

What it does

Introducing checkmate, a Robot Process Automation platform that helps to simplify yet improve the audit process by automating 3-way invoice matching. Essentially, robots go through every document and verifies each transaction.

How I built it

Using UIPath and Robot Process Automation, we automated the PDF/Excel data ingestion with an Artificial Intelligence model to extract the quantity, item description and price for the matching of the invoices.

Challenges I ran into

Difficulty in understanding and knowledge of Audit procedures and processes. Also terminology from auditors such as 3-way invoice matching are difficult to understand.

Accomplishments that I'm proud of

In the process, invoice and supporting documents are received by the robot via email. The robot opens the zip file containining documents and unzips them into the desktop. Afterwards, it opens up the individual pdfs or excel sheets and scans it thoroughly. Using AI, it matches item description, quantity and price of the supporting documents against the invoice. Subsequently, the invoice data from the supporting data is copy pasted into the relevant fields of the excel template. The robot recognizes essential information is missed from the supporting documents. If this is the case, the robot would send an email to the auditor with the request to adjust the invoice data and reliability. At the end of the process, the robot sends an email to the auditor and moves the supporting documents to the relevant subfolder.