We’ve all been on Amazon about to press buy on an item when the “Frequently Bought Together” suggestions catch our eye. An apron to complement our new BBQ set for a “fun summer day”. Or a booklight to complement the murder mystery novel for our “cozy night in”. Amazon was able to find the “theme” of our purchase and make a suggestion utilizing, you guessed it, AI.
Doc clustering is a very similar notion. It isn’t making the extra purchases for you; it’s merely grouping similar items together based on a common theme. When you submit documents for clustering, the Analytics engine determines the positions of the documents in the conceptual index. The technology takes in all “4 corners of the page” to find the concept or theme of that document. Depending on the conceptual similarity, the index identifies the most logical groupings of documents and places them into clusters. This differs from “key-word searches” that only account for those exact words, not concepts.
The current workflow we see is as follows: process the data and cull it down to a more “meaningful” set. Doc clustering is either used at limited capacity at this stage or not at all. Sometimes if there is a substantial amount of data, a client may be more inclined to use the tool, but for the most part it is underutilized and usually forgotten for the remainder of the project.
Document clustering is a powerful technique for organizing large amounts of unstructured data into meaningful groups based on their similarity. So, what’s our suggestion? Show the tool some love and use it throughout the entire lifecycle of your project.
You can run a search on the docs you want and then hop into the cluster wheel to see if there are any groupings meaningful to the case. Or with a large production it would be beneficial to run the docs through the cluster wheel and to see if there’s a theme that was missed or anything else that the case team should be aware of. If you get opposing docs why not just run them through the cluster tool and see if there’s anything interesting that pops up.
The doc clustering tool can, and should, be used to validate the discovery AND review process to ensure you are doing the job right. It’s one more good habit, one more pass through to make sure you have all the pieces of the story. Maybe the information found prompts a course correction or confirms what’s already known. From an attorney standpoint, it’s a way to be more effective and consultative to your client.
The best part is that this tool is at your fingertips, and it is a standard workflow with PLUSnxt that is included in our RelativityOne services. Once you run the analytics index it’s as simple as clicking the “Cluster View” tab within the platform. It is easily accessible, intuitive to use, and designed to make the user's experience more seamless and efficient.
In many industries, AI has become a transformative force, enabling organizations to streamline operations, enhance efficiency, and improve the decision-making process. In the context of eDiscovery, AI technologies can help organizations to prioritize and streamline various processes, such as document review, data analysis, and case strategy, thereby improving the speed and accuracy of the eDiscovery process.
Overall, the benefits of AI are numerous, and PLUSnxt’s ability to successfully integrate AI into our custom workflows has given us a significant advantage in client experience over those that do not.