Moderated by Emily, Digital Transformation Consultant at Hyperbots
Emily: Hi, everyone. This is Emily, and I’m a digital transformation consultant at Hyperbots, Inc. I’m pleased to have Shaun again with us. He is the stock compliance manager at Norfolk Southern, so great to have you on, Sean.
Shaun Walker: Absolutely, thanks for having me.
Emily: Of course. So the topic that we’d be discussing today, Shaun, is ensuring accurate GL coding of expenses. I’d like to begin by asking why it’s so important to ensure that expenses are coded correctly under the right GL codes, and what are the main challenges that companies face in this area.
Shaun Walker: I’d say it’s essential for financial reporting, budgeting, and compliance. It ensures that the financial statements reflect the true state of the company’s finances, which supports better decision-making, compliance, and alignment with regulatory bodies. The main challenge is the volume of expenses, especially in large organizations. There are so many expenses and transactions to go through, which means there’s a risk of human error.
Emily: Got it. Often, what I’ve seen is that sampling is used to check the accuracy of expense coding. Can you explain how sampling works and give examples of its advantages and limitations?
Shaun Walker: Absolutely. As I mentioned, there’s a large amount of transactions, so sampling drills down to the important ones. For example, a company might use a random sampling approach, looking at 5% of all the expense entries rather than all of them. One advantage of the sampling approach is that it’s more cost-effective and time-efficient. However, a limitation is that it may not detect systematic errors or fraud if those errors are not present in the specific transactions sampled.
Emily: Got it, understood. Also, Shaun, how do automation or automation tools in general help in ensuring correct GL coding? What are some examples of their use in companies?
Shaun Walker: Absolutely. Automation tools have predefined rules for certain expense transactions, which reduces the need for the manual process of coding and the associated errors. For example, an ERP system like SAP might automatically categorize all expenses with the heading “hotel” as travel expenses. However, some limitations can be too rigid. For instance, if there’s the word “event” in expense management, some tools might automatically categorize it as marketing, even if it was an internal training session that should have been coded differently.
Emily: Got it. Shaun, what role can AI play in improving the accuracy of expense coding? Can you provide examples of how this is implemented in practice?
Shaun Walker: One technique that AI uses is NLP, which stands for natural language processing. It analyzes transaction descriptions and suggests the most appropriate GL codes. For example, if the description is “client dinner at a restaurant,” AI would correctly code it to entertainment and expenses rather than just food and beverages. AI can also detect anomalies in real-time, helping to catch errors early and maintain accurate records.
Emily: Understood. Just out of curiosity, how can AI help in auditing human-coded expenses, and what are some examples of its effectiveness?
Shaun Walker: AI can learn from historical data to identify what correct coding should look like, comparing it to current entries. It can also flag discrepancies or unusual patterns.
Emily: Got it. Just to round things up, Shaun, one last question: are there any additional benefits of using AI over traditional sampling methods for auditing expenses?
Shaun Walker: The main thing is that humans often do it periodically, like once a month or once a quarter, whereas AI can perform continuous monitoring. AI can detect anomalies immediately, rather than waiting for periodic reviews. It can also analyze a much larger set of data, covering 100% of the expenses, whereas sampling might only cover 5% or less.
Emily: Understood. Got it. Thank you so much, Shaun, for talking to us about GL coding and how accuracy can be maintained. It was great having you, and the discussion was truly insightful. Thank you so much.
Shaun Walker: Absolutely, thanks for having me.
