Although the Text IQ tool has not officially been released by Relativity, this technology may be helpful for identifying hot spots in documents containing personal information that may need to be redacted. Its capabilities can also allow documents to be reviewed quickly after a large-scale data breach to estimate the scope of the breach and come up with a mitigation plan. It can potentially bypass a first-pass review.  

Text IQ Use Cases: 

The new Text IQ tool is supplemental to Relativity Redact. Whereas Relativity Redact searches for PII using regular expression searches, Text IQ contains 30 to 40 “out-of-the-box” detectors that are pre-determined using AI and natural language processing to identify common types of personal information. Regular expressions that can be found in Redact include contact information, credit card numbers, and personal identification numbers. Text IQ goes beyond this, detecting further information such as birth dates, addresses, PIN, social security numbers, employee ID’s, and financial account numbers. Information that you wouldn’t normally be able to catch can also be found, including information that is only partial or is misspelled or mis-formatted. You are also able to train the Text IQ model to search for certain keywords that wouldn’t usually be captured within the expressions when using Redact. 

Relativity views the major use cases for Text IQ as finding personal information, doing a privilege review, or responding to a data breach. Based on our team’s review of the technology and experience on these types of projects, we think it would best perform in the following circumstances: 

• Personal Information – Typically, PII is subject to manual redactions and requires eyes on documents. It also requires a high level of defensibility, which most attorneys view as a regular expression search. Text IQ might be more challenging in those circumstances but has the benefit of speed, so this may be best for large volume quick turn projects like a Second Request. 
• Privilege – In “hail Mary” situations, Text IQ can automate a first pass privilege review with Text IQ’s privilege scoring. I could see this being used more than the personal information feature. While Text IQ’s privilege scoring does appear to be promising, I would still want to have someone from the team review as well. This feature also overlaps with Redact but has filters such as account numbers built out with the option to add project-specific custom detectors, such as name of attorneys, law firms, etc. 
• Data Breach – Text IQ identifies personal information that has been exposed in a data breach and provides summaries to assist with notifications and reporting. I can see this feature being the most useful if you have a strict deadline for a data breach with a large volume of unstructured data. This tool could quickly gauge the scope of the data breach and how many people were affected so that you can plan your incident response accordingly. 

What I Would Like to See Implemented in the Tech 

While the Text IQ technology looks very promising, it remains in the initial stages of testing and integration with the Relativity interface. It will still go through changes as users test the product and provide their feedback. There are a few additions to the Text IQ technology that I could see Relativity implementing in the future, such as: 

• Including Larger File Sizes – There is currently an Excel file size limitation set at 75MB, however, we frequently deal with financial services clients who have documents including millions of rows and columns and would like to be sure that these files could be viewed within Text IQ.  
• Including More Languages – Currently, this tool is English-only. However, in our view, this tool is best applied to large datasets and generally the larger a dataset is, the more likely it is multi-lingual. Specifically, I could see languages like Spanish, French, and German being added first as these countries have more regulations around data privacy. 
• Eliminating Need to Convert to PDFs – The extra time needed to convert files to PDFs in Text IQ can create extra work for your team as this isn’t an automatic process in the tool. Converting to PDFs also creates incremental costs by further increasing data hosting sizes to host the PDF images of the documents. 
• Integrating into RelativityOne – The initial release of Text IQ will be on an “advanced access” basis, which may result in bugs or issues within the tech. However, this application is integrated with Relativity’s Azure environment so data will never leave your Relativity tenant. 
• Ensuring Tool Isn’t Overly-Inclusive – Similar to Relativity Active Learning, we’d like to see reporting capabilities to validate that the tool is not over-inclusive or under-inclusive. Ideally, we’d also like to see statistics from Relativity on that topic as well. 

If you would like to see a demo of Text IQ, please feel free to reach out to me at Josh.Treat@AcornLS.com. I am also happy to discuss how this technology may work for your case, or exchange thoughts on what other technology in the market may fit.