Challenges with accurately categorizing line items for tax purposes

Find out interesting insights with Dave Sackett, VP of finance at Persimmon Technologies

Moderated by Emily, Digital Transformation Consultant at Hyperbots

Don’t want to watch a video? Read the interview transcript below.

Emily: Hey, everyone, this is Emily, and I’m a digital transformation consultant with Hyperbots. Really pleased to have Dave Sackett on the call with me, who is the VP of finance at Persimmon Technologies. Thank you so much for joining us, Dave, today.

Dave Sackett: Yeah. Thanks. Emily.

Emily: So, Dave, the topic that we’d be discussing today is challenges with accurately categorizing line items for tax purposes. And I’d like to, you know, kick things off by asking what are the main challenges with accurately categorizing line items for tax purposes.

Dave Sackett: I’d say the primary challenge is that line items often lack sufficient description, making it difficult to determine the correct tax category. Sometimes only a part, number, or generic term is provided without a specific line. Items are determined to be taxable or not. So, for example, an invoice might just say hardware without further details, leaving ambiguity around. Whether it’s taxable. Hyperbots can address this with AI technology that identifies tax-relevant data based on limited descriptions or part numbers.

Emily: Got it. Now that you mentioned AI technology, Dave, how does Hyperbots AI manage to categorize items with limited descriptions like a part number, or you know, let’s say a generic term?

Dave Sackett: Yeah, no. AI technology is constantly advancing Hyperbots. AI uses machine learning and natural language processing to analyze context clues in the invoice. So, for example, if an invoice lists, only a part number Hyperbots can cross-reference it to an external database to identify its category. If the part number matches a known item. This standard, such as a standard hardware, component Hyperbots, categorizes it by leveraging its extensive database of past categorization, and this means the AI can leverage both internal information and external information to make that determination.

Emily: Got it. Can you also expand a little bit on how Hyperbots use context from the entire invoice or purchase order to improve categorization and accuracy?

Dave Sackett: Sure. Yeah, Hyperbots examine surrounding items on the invoice to establish the context of where that information is coming from. So, for example, if a line item says license. But then other items include the words software or support, the system infers that it relates to a digital project or service, guiding it to the right tax category. This context-driven approach helps categorize even vague terms by using information from related terms to form the complete picture. This is a newer AI technology that can help identify this and now be used in a way to help with categorization.

Emily: Understood. And what data sources do Hyperbots leverage to enrich limited lines? Item, descriptions.

Dave Sackett: Yep. Hyperbots enrich data by tapping into prior purchase history vendor records, public part databases, and other information available on similar items. So, for example, if AI encounters a generic part number from a vendor, it checks historical purchases or industry databases to identify its category, ensuring high accuracy in categorization. This approach minimizes the risk of errors due to incomplete descriptions.

Emily: Got it talking a little about compliance, Dave. So how do Hyperbots ensure compliance with state-specific tax requirements when categorizing items?

Dave Sackett: Yeah, Hyperbots integrates a comprehensive state-specific tax category database updating it continuously to reflect the tax changes in state laws. So, for example, if an AI system knows that software licenses might be taxed differently in California versus New York and applies the correct categorization based on the state requirements. This ensures compliance and eliminates the risk of applying incorrect tax rates across different jurisdictions. And if you were to be audited, there’d be that audit trail that AI determined the right tax based on the right methodology.

Emily: Understood. And what do you know, what data does Hyperbots rely on in order to achieve 100% accurate tax categorization?

Dave Sackett: To achieve high accuracy. Hyperbots rely on several key data points, detailed item descriptions, part numbers with historical records, vendor details, transaction history, and state-specific tax categories. So, for example, if it has a full description like a digital software license, it categorizes it quickly and accurately. When these details are combined, the AI ensures precise tax calculations for each line. Item.

Emily: Got it, Dave, can you talk a little more about it? You know, how does Hyperbots’ proprietary technology continuously improve its categorization? Accuracy.

Dave Sackett: Yeah, Hyperbots. AI incorporates feedback from past categorization corrections to improve its future performance. If AI categorizes the line it’s later adjusted by the user. Hyperbot learns from this adjustment and applies new knowledge going forward. This is reinforcement learning that ensures that the AI constantly evolves adapting it to changes in product offerings, tax rules, or categorization standards. So the more you use Hyperbots, AI, the more it learns, the better it is to do its job with the correct training.

Emily: Understood. And just one last question, in order to summarize everything right? Can you summarize the primary benefits of using Hyperbots? Proprietary technology for line item tax categorization.

Dave Sackett: Sure Hyperbots provides a fully automated, accurate categorization solution that minimizes manual intervention and ensures compliance with state-specific rules. For example, by using context, clues, and part numbers. Hyperbots can categorize items accurately, even with minimal descriptions, saving time and reducing errors, it continuously learns the capability further, improving accuracy, and making it a scalable solution for high transaction volumes across multiple locations. So that automation of the tax categorization is going to save time and by improving accuracy that also saves time.

Emily: Got it. Thank you so much, Dave, for you know, expanding more on Hyperbots and talking through the challenges with accurately categorizing line items for tax purposes. It was a truly insightful discussion. So thank you so much for being here.

Dave Sackett: Yeah, thanks, Emily. Great to be here