Thanks, Greg. I'm Aki Founder and CEO of Genial
Technology. We make audit easy and intuitive with AI based software, so that
busy auditors can focus on the area where human judgment and experience are
required. Have you been reached out by an auditor?
Okay, some of you have. They are annoying right? Auditors ask tons of questions.
Auditors need to test sample transactions and check if they're okay.
Accounting teams have to prepare many documents for audit purposes. These
works will be easier and faster with our software. External audit's workload has
increased over the last 10 years. But, audit technology improved very slowly.
Manual audit is slow. Excel is not optimized for audit. It's also labor
intensive to check massive amounts of documents. Moreover, it's easy to make
mistakes, and there are several layers of reviews to prevent such errors. So, Manual
audit has high risk and high cost. The total audit fee is around 40 billion
dollars in the North America. That's why we have developed GenialAI.
GenialAI is an online audit software based on AI technology. Alright, let's
jump into the demo. In this demo, I'll show you how GenialAI looks for frauds
and errors from ERP data with PDF attachments and how it shows the
details. We are running a SAP instance in our development environment.
This sample transaction data has typical feels like journal number, dates and
amounts. Also, there are bunch of PDFs attached to each transaction, such as
credit card logs and expense reports. Before doing the test, auditors don't know where
issues are. So, let's export it to GenialAI.
Our SAP plugin software automatically copies the data to GenialAI. Then, I
switch to GenialAI and login to the system. It is important and processing
the data with PDFs from SAP... I'm sorry. Let me try this again. Now,
GenialAI job is running to identify frauds and errors. Example of frauds would
be double dipping, fictitious sales, bill padding and duplicate payments. Errors
would be data entry mistakes. It may take more time to process all the data, so
let's look at the previous result. Can you see it? GenialAI identified potential frauds
are the pink cells and errors as the yellow cells. It's done. It's simple and
fast. That's all. But, it's too simple for demo, so let's look at the fraud case,
double dipping. Double dipping is a type of a duplicate
expense claim where one is through corporate card and the other through
expense report. For instance, GenialAI found these two lines to be a double
dipping. The amounts are both $300, and the dates are not far from each other.
They are posted on July 3rd and June 30tn. This is also identified based on the
OCR results of the PDF attachments. GenialAI automatically downloads the
attachment files, so accounting teams do not have to email or print them out for
auditors. For instance, this credit card log, attached to one transaction,
indicates that Samara Johnson spent $300 at a restaurant on June 30th. On the
other hand, this expense report attached to the other, shows she spent
the same amount at the same place on the same day again. Therefore, GenialAI
thought that she embezzled $300 from the company. This is double
dipping. I hope there is no person here named Samara Johnson in this room. Okay,
so this is how GenialAI finds frauds and errors. How to configure these processes? Our
demo ruleset consists of these rules, and you can customize them by adding or
deleting the rules and fields. OCR tokenizer reads PDF supporting documents.
Pre-process like whitespace removal, and find frauds like double dipping. This is
how GenialAI tests frauds and errors from transaction records. On top of this
transaction testing module, now we're developing sales testing module by the
end of this summer. This is another major area where regulators have a lot of
focus on. Auditors are required to test hundreds of revenue transactions with
supporting documents, such as purchase orders, invoices and shipping documents,
with their eyes, and there is too much back-and-forth communication with their
client counterpart. So, current process takes average 250 hours per client per
year in sales testing. GenialAI can save it to average 50 hours by automating this process.
Additionally, GenialAI can test the entire sales data set, instead of samples, which
is impossible even with thousands of hours of human resources.
Furthermore, we are conducting a pilot test with AI Audit Research Lab at one
of the Big Four accounting firms to measure performance of GenialAI. They
will provide sales data used in actual audit projects and try GenialAI. So,
my main message is that GenialAI will allow auditors to focus on the
higher-risk areas where human judgment and experience are required. If you're a CPA
who works on audit or you are in an organization having same kinds of issues,
please come to my booth. Thank you
Không có nhận xét nào:
Đăng nhận xét