The old way of capturing essential data attributes from legal documents has been transcended, thanks to technology. we hear about Artificial Intelligence (AI), we hear about Machine Learning software systems and other new technologies for data extraction from contracts. But when we dig into the details, the accuracy of abstraction is not to the standard for companies to make reliable business decisions. It is difficult to achieve very high accuracy from software and consistency from manual methods.
The systems cannot adapt to the full variability of documents from which the essential data elements are to be extracted. There may be errors due to OCR – the process to convert image-based documents into text-based documents (a “5” may be converted to an “s”), causing any software to either miss or wrongly extract the data. Furthermore, there may be handwritten attributes.
So, software alone cannot be the answer. Even with the claims of extracting 90% of the information from documents, when digging deeper, it is not the same 10% that is incorrect. And 90% extraction is only under ideal conditions – such as perfect OCR, no missing information in contracts, etc. In a contract management situation where risk and rewards are so prevalent, and critical, anything less than 99.99% is not acceptable. The attributes you miss may include the one that carries the penalty or has the reward.
So, going with a hybrid solution i.e. a combination of both software and human will a viable answer to all your problems. Here the software can bring speed to the process and human intervention can deliver required accuracy.