Inspiration
A lot of manpower is spent looking through invoices and receipts, recording them down into the system, then running analysis on the number in order to find discrepancies. This restricts overall scalability and efficiency of audit checks.
What it does
EZAudit converts every copy of the receipt into a digital copy using Optical Character Recognition (OCR), stores it online, extracts key information such as ‘Company ID’, ‘Receipt Number’, ‘Final Amount’, and ‘Date & Time’ out into Microsoft Excel CSV format, and compares it against the host company’s ledger for discrepancies. The Auditors can then spend their time to investigate the discrepancies instead of manually keying such information. This allows the Auditors to scale up their work, take a larger sample or reduce hours spent on low-level tasks. The video we have shown is a "rough concept" of how the OCR will work. In the actual scenario, a mobile app will be developed to take a picture of the receipt and send it to the database where the conversion, extract and analysis will take place, and discrepancies are flagged out in the system and displayed in the Auditors' computer (web app) for them to investigate. Please download the zipped file and slides from the Google drive link below.
How we built it
In terms of technology, we used Tesseract & C#. Google Cloud is used as the backend to provide the most suitable platform for OCR development. Also, 100% of it came from hard work & grit.
Challenges we ran into
Our team did not have an extensive understanding of both OCR & the key challenges in accounting audit. Thus we had to actively reach out to experts in respective subject domains to better understand both the audit workflow and the key technologies to consider in our prototype.
Accomplishments that we're proud of
We were able to quickly identify the daily problems of Auditors and designed a solution that makes auditing easier. We turned something that was not easily scalable into one that was. Rather than implementing technologies that were in the development phase within the short time frame, we made use of existing technology. Thus we did not have to completely reinvent the wheel to innovatively address the extremely manual tasks of auditors and enable them to perform higher level tasks.
What we learned
The audit industry is a passive industry that relies heavily on the client’s accounting system in order to make their audit job simpler. We realised that there is a lot of potentials for auditors to better harness technology to improve their workflow and efficiency given the current legal & technological landscape in Auditing.
What's next for Boulderaze
Every receipt and invoice comes in different format sizes. The next step of our product development is to train the tool to recognise handwritten forms and other variation type of receipts and invoices, extract them out, sort them, compare them against the ledger, and automatically flag out the discrepancy found. We will look into addressing these issues using machine learning technology. We will also focus on implementing TAGUI to further automate the process of consolidating receipts/invoices.
Log in or sign up for Devpost to join the conversation.