November 18, 2022

Innovation Insights Podcast: Trullion

By: Center For Accounting Transformation / podcast
Image

Leveraging AI in Accounting with OCR and NLP

Innovation Insights host Donny Shimamoto, CPA, CITP, CGMA, welcomes Isaac Heller, the co-founder and CEO of Trullion, to discuss the origin of technology pioneered by the accounting and legal professions that continues to grow and allows automation of mundane tasks.

AI uses optical character recognition (OCR) and natural language procession (NLP) to read documents, like PDFs and Excel spreadsheets, and extracts relevant data, like lease dates, into databases, where they can be used for computations or searching.

“When you think of Oracle and SAP and Intuit going way back, these were really the pioneers of database software,” Heller says. “It’s as simple as that, you know, the general ledger and the accounts payable and accounts receivable framework fits really, really well in a software model. So, there was kind of this jump out of the gate where we were able to do so much in accounting.”

Shimamoto and Heller share how accounting was the perfect breeding ground for database software and how it’s the perfect field to direct smarter technologies for the future.

Key Takeaways:

  • Because you’re starting with the source document (e.g. PDF, Excel) and extracting the data from there to then put the data into the system, you can trace data back all the way to the source document all within the same system (not just to the point where someone keyed the data into the system).
  • The goal isn’t for the machines (AI) to do everything; it’s for it instead to complete 70-80% of the grunt work and allow the accountant or auditor to focus on the analytics and strategic work.
  • Small and regional firms can gain an advantage by partnering with accounting technology vendors that enable them to innovate and incorporate technology into their services more quickly than larger firms.

Read more about Isaac Heller and Trullion at CPA Trendlines.

 

Subscribe:

 

Share This Article