How AI is Revolutionizing Land Workflows

ChatGPT, Microsoft Copilot, and Google Gemini can pull off astounding feats in certain contexts but in many ways generative AI epically fails. In a recent project looking for research on title defects in oil & gas leases, ChatGPT invented convincing studies that never existed and stats that came from no reliable source. If you've explored generative AI you've probably run across a similar hallucination where the technology appears to be telling you what you want to hear. In machine learning, this is called ungrounded content and in an industry where the details of a lease determine the destiny of an operator, generative AI needs careful watching.

While chat bots and GenAI have dominated headlines, other branches of artificial intelligence have been quietly innovating for a decade and the good news for the land department is that solutions like Thomson Reuters Document Intelligence (formerly ThoughtTrace) and BIS Grooper are designed with leases and complex contracts in mind. These solutions utilize natural language processing (NLP) and have been trained with machine learning algorithms to extract data from the most difficult to read documents.

Optical character recognition technology has been around for decades with tools like Adobe Acrobat and Google Docs able to OCR files in real time. OCR comes with a 2% character recognition error rate if you're lucky and have pristine paper documents with clear and standard fonts. But as anyone in the land department can attest, lease records from just a few years ago can suffer from readability issues, not to mention records that date back decades or even to sovereignty.

NLP products combine computer vision to bring a tremendous level of accuracy and automation to not just OCR land files but understand context and add structure to unstructured data. For example, a PDF file of a scanned lease packet often contains a dozen or more pages. Rather than managing files and pages however, computer vision enabled NLP solutions find the business objects that matter most to land professionals, including key lease terms like Pugh and Mother Hubbard clauses, general obligations, royalty clauses, and expirations. This data can then be loaded directly into your land management software.

Fortune 500 CIOs and leading IT researchers like IDC and Gartner have frequently argued that if your team does not have a strategy to leverage AI in 2024 then you're missing a once in a lifetime, transformational opportunity akin to the industrial revolution and the advent of the Internet. While hype around GenAI persists, NLP is a proven technology that is available now and deployed at many of your peers. Pepper Energy Partners is a certified partner for Thomson Reuters Document Intelligence to help you transform vast volumes of land files into valuable data.

Learn more or contact Pepper today.