How to Streamline Disclosure Process with Technology Assisted Review (TAR)?
Last updated:
07 Sep, 2023
By Ritu Kaushal |
5 Minutes Read
incur Technology is taking over the front seat in every sphere of life, and the legal profession is no exception. One of the areas that courts and tribunals encourage is technology-assisted review (TAR). It is particularly helpful when analyzing large volumes of documents in complex cases.
However, you need to know that TAR can automate the document review process. However, you need to monitor the process to optimize the effectiveness of TAR and ensure that no slip-ups occur at any stage.
What is TAR?
TAR is a broad term that denotes any tool or approach to document review. Any technology that facilitates the legal document and management review process and makes it more efficient and reliable comes under TAR. It eases the process and is suitable for cases where you have to analyze and review voluminous amounts of documents. Even in small cases, you can use TAR to serialize or cluster documents to present to human reviewers in an orderly fashion.
The gains from TAR are usually more conspicuous in instances where the number of documents under review is typically greater than a hundred thousand. The potential efficiencies should match the operational costs of the process. This is not a standard process because every litigation case differs.
The process of TAR includes the following concepts and activities:
- Predictive coding–
Statistical models rely on definite algorithms that you may utilize to recall documents similar to a sample set. This sample set is a “seed set.” A reviewer then verifies whether the documents are consistent with the seed set. Once checked, the documents are again fed back into the algorithm, thereby increasing the size of the “seed set.” This form of predictive coding can’t work as a substitute for subjective document review as the latter has to to ensure consistency and integrity of results.
- Concept Cluttering–
This term denotes document sets or search results that are in an interactive and graphical format. TAR ensures that documents with similar content are in cluster form, even though the display configurations may vary between legal back-office service providers.
- Email threading–
This is the process where you can use statistical methods to link up emails that are similar in nature but aren’t identical to each other. Examples are email chains that branch off in content from a main email thread or emails that you forward to different recipients.
- Identification of near-duplicates–
Like email threading, you can utilize identify and club together documents that you don’t consider identical but have similar concepts. For this TAR uses “fuzzy” searching, synonyms and custom dictionaries to arrive at a common plank in sorting documents.
How to optimize the use of TAR?
While the expertise of document review services is never in question, optimization in using TAR to its fullest potential depends on following a set of guidelines. Here are a few of them:
- Constant monitoring and vigilance–
TAR is not an automated process, and constant monitoring is necessary to get precise results. The legal back office team has to decide critically which documents they require for subjective review. Documents that you recall review them manually on a sample basis to ensure that the system is working properly. Once you are through with corrections, the algorithm will come back with the modifications.
- Reduce the number of reviewing lawyers–
Unlike conventional document review, where lawyers engage, TAR achieves greater efficiencies as only one or two experts review a small sample set to check the accuracy. Once done, the algorithm will do the main work of identifying related documents, which a relatively junior team can then check. It is important that the initial work of checking the smaller training set of documents to a senior lawyer only. On average, a sample set of as few as 3,000 random documents may be verified for one million documents. That is the power of TAR if used judiciously.
- Run multiple TAR processes for disputed issues–
TAR processes do not work optimally if there are a number of issues or lots of categories for discovery. In such cases, perform separate TAR procedures in each case with lawyers who are experts in different aspects and supervise them. There is a disadvantage to running multiple TAR processes because it is difficult to avoid “double-handling,” where you review documents more than once for different purposes.
- Create “example documents”–
The legal team may create “example documents” and then let the algorithm find similar documents. You can accomplish this by testing whether the system is working efficiently or not. These documents should work as a testing tool for email threading and other cluster tools. Quarantine the dummy document after testing the disclosure processes.
- TAR is not infallible and should be treated as such–
While TAR has great benefits, it is not infallible. It identifies documents by the content in them only and not what is excluded. But that too can be used to advantage. A fraud if concealed in a daily report intentionally over time will also be pulled up in TAR. The fact that it is not there in all documents will establish the matter of intentional concealment. Similarly, TAR does not work on photographs, drawings, and spreadsheets.
Follow these guidelines to get the best out of TAR to streamline the document review process.
Are legal document review processes proving to be troublesome to handle? Consider outsourcing the same to Cogneesol! We provide legal document review services, ensuring optimum quality. Call us today at +1 646-688-2821 or email at [email protected].