Lawyers are just starting to get comfortable with the use of Predictive Coding to increase productivity and reduce the overall cost of document review. However, as I have been predicting for almost 24 months, Predictive Coding in the form of Predictive Analytics or more correctly called Machine Learning is now beginning to show up as a very effective way to identify documents for Records Management and other general Information Governance requirements.
On November 20, 2013, Ben Cole, Site Editor for SearchCIO.com, published an excellent interview with Leigh Isaacs, Director of Records and Information Governance at Orrick, Herrington & Sutcliffe, LLP, that covers how Predictive Analytics is evolving.
The full text of that interview is as follows:
Predictive coding is traditionally associated with e-discovery strategy, but companies are beginning to see how the technology and its associated tools can benefit everyday information governance. Leigh Isaacs, director of records and information governance atOrrick, Herrington & Sutcliffe, LLP, said innovative use of predictive coding technology can help with information management, increase efficiency, mitigate risk, and maintain an organization’s competitive edge. Isaacs led a session titled, Predictive coding: Leveraging analytics technology for information governance, during a recent SearchCompliance virtual trade show, and after the presentation, sat down for the following Q&A.
How is predictive coding technology adapting to the increased use of mobile devices?
Leigh Isaacs: This will likely be an ongoing, evolving topic. Typically, we recommend organizations ensure any company information that might be on a mobile device is synced with the organization’s internal systems. From a records retention and e-discoverystandpoint, it is best practice not to allow unique data to reside solely on mobile devices. Addressing this, I think, is more of a policy/process issue at this time.
What are some of the key questions to ask predictive coding vendors, and what type of info should be included in the service-level agreement (SLA)?
Isaacs: I would first recommend the organization identify their objectives and the issues that they want the solution to [address]. Develop your business and system requirements. Since using predictive coding for information governance is new, I think the vendors need as much information from us as we do from them. There are also varying things we want to accomplish. The questions will be different if you want to use predictive coding for data remediation than they will be for information management and maintenance.
I’d ask about the pricing model. I’d ask about data retention and storage, how their system works. Can it natively work across company repositories? Or, do you need to ingest the data into their system for analytics, which is likely in the cloud? If so, I’d include in a service-level agreement anything that would typically be found in a cloud-related SLA. I’d also be sure to work closely with general counsel and/or e-discovery counsel to develop the requirements and identify the questions that need to be asked.
Can predictive coding technology help with data collection efforts to maintain regulatory compliance? If so, how?
Isaacs: Yes. By using predictive coding technology, you can better identify where sensitive information, such as PII[personally identifiable information], is, and make sure it is appropriately managed and secured. By using predictive coding technologies, coupled with a well-developed predictive coding process, you can maintain an audit trail of how this information was generated, used, maintained and secured.
Who should champion the use of predictive coding technology in a company? Is that the legal department’s concern, or is it more on the records management side?
Isaacs: This is an area that is quickly evolving. Much like information governance in general, the person who should be carrying the flag to champion predictive coding varies. If records managers have a real interest in this technology, which they should, then they can certainly be the ones to rattle the cages to get some attention. That being said, going forward with a plan needs to be coordinated and communicated between records management, legal and compliance. Again, it depends on what you want it to do. I often suggest finding some low-hanging fruit to make some progress, and then raise the bar further with actual metrics and ROI of usage.
Do trends such as the cloud, social media and mobile devices complicate predictive coding efforts for organizations? If so, how can companies overcome these obstacles?
Isaacs: Cloud, social media and mobile devices complicate everything for organizations. My answer will sound simple, but it’s not. As I mentioned, I think it really comes down to a combination of people, process and technology. Technology is great. Predictive coding is great. But tools alone are not enough to solve our problems. If possible, there needs to be a proactive approach between all company stakeholders. Decisions need to be made and communicated about expectations around information governance and how technologies will be used and monitored. Again, predictive coding is great, but it’s still not a magic wand solution. Like everything else, the capabilities need to be factored into the overall strategywhen developing a predictive coding program.
- How to Engage Enterprise Buyers in Meaningful Conversations in 2016 February 28, 2016
- nVIDIA Driving Deep Learning to the Forefront – Literally February 22, 2016
- New Technologies Disrupting the Legal Business in the UK February 17, 2016
- Shares of Tableau plunge 36% after company posts $41M loss in Q4 February 5, 2016
- LexisNexis Unveils Lexis® DiscoveryIQ eDiscovery Platform Enhanced by Brainspace February 2, 2016