Conceptual Search And Auto-Categorization Technology: The Latest Step In The E-Discovery Process

Saturday, March 1, 2008 - 01:00

Editor: Mr. Carpenter, would you tell our readers something about your professional experience?

Carpenter: I began my career on the law firm side and joined the vendor side after several years. As general counsel at Recommind I continue to be engaged in the practice of law.

In the broad sense, my job is to take stock of the various tools that are available in the market to support attorneys in their work. Some of the technologies that are available today are simply mind-boggling. I wish I had had access to them during my law firm days. Things were much simpler then, of course. In the discovery process, the documentation to be reviewed was finite, and the process itself both easier and quicker. And, for the most part, cost-effective. Today attorneys have to worry about a much broader range and scope of documentation. Did the client find everything? Did they look in the right place? What format are the documents in? The tools that attorneys utilize to get them through this process efficiently do not constitute a luxury, but rather add up to one of the absolutely essential components of effective litigation.

Editor: Can you give us an overview of Recommind?

Carpenter: Recommind has been in existence for about 10 years. It is, essentially, a conceptual search and auto-categorization technology organization. We possess one of the very few technologies that permits us to look at documents, automatically extract concepts from within and then relate those concepts to similar concepts across an entire universe of documents. At the same time we can categorize documents based on their content. This constitutes our core technology, which we apply in several ways. Initially we focused on the knowledge management market and served as the enterprise search solution for large entities, namely large law firms or the legal departments of large corporations. We continue to do that very well and are by far the market leader in this particular arena. Within the last 18 months we have branched out using this technology to enter the e-discovery and e-mail management areas. That is, we use our conceptual search and auto-categorization technology to address the review and analysis phase of the discovery process and to address e-mail management for the client's higher-end knowledge workers or attorneys.

Editor: What can firms do to improve the document review process?

Carpenter: The market has done a good job of addressing the processing phase in terms of reducing the overall volume of documents. The focus is now on the review phase, where some 70 percent of the costs are incurred. Increasingly, people are using our technology - and specifically auto-categorization and auto-coding - to do a first pass review. This entails a minor bit of coding of certain documents for priority and relevance on the part of an attorney. The system then extrapolates the wording from those documents across the universe of documents.

Now that the entire document collection has been pre-coded, the actual attorney review can be organized in a number of different ways. By subject matter, of course, but also by a presumption of privilege, which allows the documents to be put in front of the right people from the outset of the review. Another important feature of the process is that it gives the individual reviewer an incredible amount of context. The document comes pre-coded with essentially a confidence score. This does not replace the job being done by the reviewer, but it does augment and dramatically speed up what he is doing while improving the consistency and quality of his work. And, as I mentioned, it is at this stage that most of the costs are incurred.

Editor: Can you give us some examples of how an organization saves money by recourse to this technology?

Carpenter: The most obvious example is where the reviewer can take 30 seconds to review a 25-page document as opposed to taking five minutes. There is a considerable savings where the reviewer is able to look at a computer-generated guess as to how the documents should be coded and then eyeball the key terms that are highlighted, rather than reading through all 25 pages with no context.

With the revised Federal Rules of Civil Procedure coming into effect in December of 2006 the process of pre-trial conferring has changed. Prior to that date it was not necessary to know, going into a conference, the contents of the document collection. Today, the attorney who knows whether he has good documents or bad at this stage has a distinct advantage. The technology permits him to get the documents in front of the essential reviewers quickly and efficiently, and that is going to save money and, very probably, have a dramatic effect on how the case turns out.

Finally, the technology gives the attorney quality control over the process. I continue to be astonished at how many attorneys do not have any way of measuring the effectiveness of the review. We provide that measure.

Editor: Is technology ready to assist humans in the review process?

Carpenter: Yes, absolutely. It is not ready to replace the human component, but it is a major step forward. Conceptual search, for example, permits me to enter a query for, say, shipyards and receive back all documents with that term and synonymous ones automatically. That was not possible in the past with keyword search.

The new technology is also important where there are multiple parties in a case. A single key word for all parties will find synonymous materials, as opposed to key words for each party.

Editor: Are these technologies defensible in court?

Carpenter: No one size fits all. Most litigators, in my experience, are not comfortable if they have failed to review the documents themselves. That means, very often, that they are not comfortable defending a review process in court which has been outsourced or through which the documents have been sent through a purely technological process and coded.

What we vendors are attempting to do is speed up the process and get the right documents into the hands of the attorneys efficiently and quickly. It is critical that they, the attorneys, are the ones to make the coding decisions because that is defensible in court and something that judges are increasingly willing to accept.

Editor: What preventative measures can corporations take to prepare for litigation?

Carpenter: There are three measures that appear to be of great interest to our customers today. Of particular demand is expedited review, some kind of computer-generated or augmented review. A second area of great interest is information management for preparation of litigation. And finally, legal holds.

Information management in anticipation of litigation is particularly important. We hear from many of the top tier law firms and large corporate law departments that their principal concern has to do with the effective management of information. Knowing where the information resides and how it is maintained is the first challenge. A second has to do with its organization: what are the relevant subsets of information? Is it kept for business purposes? Compliance? Records management? E-discovery issues? An effective organization of the information permits the firm or company to meet all of its documents retention requirements without getting in the way of secure information access.

Editor: How can technology help when a corporation implements a legal hold?

Carpenter: The information management perspective is relevant here. The most important reason people must implement a legal hold on their information is for e-discovery purposes. When this process first began, the company would send an employee to a site and have him copy documents to some sort of shared drive. The next step was to utilize technology to go over a network connection and take a snapshot of the entire drive. This is, in fact, the current standard. However, no enterprise has an interest in keeping everything , and with litigation once the material is imaged it is very difficult to eliminate. What is increasingly desired is the ability to use technology to reach into each desktop or laptop computer, run a query locally and only pull back the bare minimum of information needed for the legal hold process in order for the enterprise to defend itself in court. This is the future of legal holds. The technology I have described will be with us in about six months.

Editor: Are there other uses for legal hold technologies?

Carpenter: Yes, there are quite a few. One good recent example concerns the European investment bank Société Générale, which has incurred trading losses of some five billion dollars as the result of the activities of a rogue trader, Jerome Kerviel. The bank might well desire to keep certain information under wraps concerning what their traders were doing prior to the series of events that led to the losses. It would be helpful to run periodic inquiries or simply track certain information using the same technology that we use in a legal hold scenario. That way they can preserve the information on hand whether or not there is a regulatory inquiry, investigation or lawsuit underway.

Editor: How can litigators use the revised FRCP and new technologies to their advantage?

Carpenter: For the average litigator the revised Federal Rules present a great opportunity to gain a strategic advantage over the other side. They, the litigators, must become fluent in the language of PST files and native versus TIFF format. But once they have some fluency in these languages they will know where to look and how to drill down into their document collection, and they will be able to do this at an early enough point to understand the strength of their case before the other side does. This permits them to manage the gamesmanship required in litigation very effectively, and precisely the same strategy prevails when negotiating pretrial settlements. If you understand your case and what you can and cannot show to the other side, you have a much better chance of coming out on top than if you are trying to forge ahead on educated guesswork.

Editor: Is there anything you would like to add?

Carpenter: We are in a very interesting time. The legal industry is going through considerable disruption, and much of it has to do with the exponential increase in the volume of information. The technology that permits attorneys to stay abreast of this development - even ahead of it - is there and it is being enhanced on an ongoing basis. Those attorneys who take advantage of technology, who speak its language and exploit its benefits, are going to be a long way ahead of those who do not.

Please email the interviewee at craig.carpenter@recommind.com with questions about this interview.