From the legal industry to the insurance industry, a considerable amount of time and money is spent reviewing and tagging documents. Well, that’s actually the politically correct way of putting it. In reality, a massive amount of time and money is nonnegotiably wasted reviewing and tagging documents.
The good news is, there’s technology called predictive tagging (or predictive coding) that finds responsive electronically stored information (ESI) documents during any type of review or search process. If you’re in the legal industry and in the legal case’s review phase, this predictive tagging/predictive coding feature is called Technology Assisted Review (TAR) or computer-assisted review (CAR). Predictive tagging uses artificial intelligence, namely neural networks and deep learning, to develop software that continues to learn and make better decisions while significantly expediting the review process, saving time and money.
Whether you’re in the legal industry or any industry where finding and tagging documents is necessary, this predictive tagging can help you. Basically, predictive tagging (or takes the human out of the loop and empowers a machine to do the grunt work. But we're also not singularly letting a machine decide how to review and code a document. Anvesa learns from your actions and classifies, filters, and tags documents for you so you can focus on broader issues, like strategizing on winning a legal case or finding key information needed to file an insurance claim.
Predictive tagging starts by training a pre-defined model with the help of a seed set of data—a sample of documents pulled from the entire group of reviewed documents. Reviewers code each reviewed document as responsive (relevant to the case) or unresponsive (not relevant). If the input seed set is poorly reviewed, the result will not be satisfactory. The predictive model will fail to train, and more reviewed documents will be required to train the model successfully. The more accurate the tagging of the seed set, the more accurate the machine learning and predictive tagging will be. As the training continues in form of Active continuous learning (ACL), artificial intelligence enables the software to learn and make better, faster decisions as time goes on.
If the training is successful, the software is even capable of bulk tagging documents by types with class probability. Documents are tagged as part of suggestion to the user with different degree of probability I.e., “very likely”, “likely”, ”less likely”, and “unlikely”. The higher the probability, the stronger the suggestion is.
Also, the great thing about Aureus’s predictive tagging/coding technology is that it’s extremely cost-effective, as showcased by Aureus’s eDiscovery review product Anvesa®. The user only pays for the space they need.
And while predictive tagging is a hot buzzword in the industry, it has yet has not been widely adopted into everyday practices. Like with any new technology, there is an adoption curve. However, predictive tagging is commonly accepted by judges in the use of e-discovery and is expected to become even more frequent as the use of predictive tagging expands. So, the real question is, if it saves you time and money, what’s stopping you from using it at your company?