Global menu

Our global pages

Close

Eversheds Sutherland’s LitTech Revolution

  • United Kingdom
  • Litigation and dispute management

05-03-2020

How Propel is changing our Litigation and Disputes Practice

Data is a source of interest, debate and discussion. It is undeniable that our lives and overall digital presence in the professional and personal world are the rule and putting pen to paper the exception. A variety of legal cases carry “electronic disclosure” as a key part of their overall lifecycle, and data collection is one of the first steps to consider alongside a range of forensic technologies and expertise. Examples of data sources falling within the scope of collection include:

  • Email data, by way of collecting individuals' entire mailboxes
  • Loose files stored on electronic devices
  • Data stored in a cloud, such as Dropbox, Google Drive etc.
  • Data stored on shared drives
  • Files from central repository systems such as document management systems, financial reporting systems and banking systems
  • Hard copy documents, later digitised by way of manually scanning and coding them to produce an electronic version
  • Mobile device data

The scope of each investigation will capture all or some of these categories, and the intervention of forensic experts is required in order to ascertain the defensibility of this data.

Times were simpler when one merely needed to drag and drop files on a “USB stick” and that would be classified as a data harvesting exercise. However, given how big data has become into everyday use and the multiple locations of storage, data collection is an important yet complex exercise.

The range of data collected for analysis and legal review will always define the legal teams' decision-making process. An entire case strategy is built on data and the analysis tools available to the teams, which also play a huge part in what comes next. Mining data is merely the beginning and it will almost always be followed by:

  • Data Culling
  • Data Analysis
  • Data Review, and ultimately
  • Data Disclosure

Data Culling and Analysis are frequently part of what we typically would call the “Early Case Assessment” phase of a case. It is of paramount importance, and a phase that requires consultative guidance by eDiscovery experts. From adopting simple combinations of data ranges and carefully selected search terms to structured analytics, such as domain parsing, and language identification or conceptual analytics, such as phrase analysis, clients can get to the heart of the issues in hand much faster and identify immediate issues with the data (for example, a gap analysis of the data and potential periods of time that data is missing from the initial collections).

Technology can assist in a plethora of ways and at every stage of an investigation and/or a disclosure exercise. Forensic tools specifically designed for data collection apply a unique methodology of “capturing” the data not just as it is (present image of the data on a computer, laptop, phone), but also digging deeper in archived data, deleted data, back-ups of data, even in the abstract cloud storage locations, and deliver what is considered an accurate “image” of all data belonging to an individual or an entire company. Tools such as EnCase, FTK, Cellebrite are just a few of the products on the market that meet the requirements of data collection, including imaging of mobile phone devices and capturing of chat messages.

Moving from data collection to the actual identification of what the data holds within and the analysis of it all is where Eversheds Sutherland’s Litigation Technology team, known as Propel and based in the United Kingdom, can offer a wide range of services supported by the tools we hold in-house.

ES Locate (powered by Axcelerate) is our electronic discovery, review and analysis platform that is hosted in a secure private Cloud location. Amongst its various capabilities, here are a few worthy of highlighting:

Search terms

After data sets have been processed, ES Locate immediately applies simple or sophisticated combinations of search terms to cull the data set into more manageable document sets for review and analysis. We can revisit the search terms as many times as needed, providing clients with guidance on search syntax and how best to find what they are targeting. ES Locate also provides detailed search terms reports that allow clients to see term hits, which search terms are unique, and document counts with full families, so clients can assess their review populations.

Advanced data analytics

Phrase analysis is part of Conceptual Analytics and allows clients to find how terms and phrases appear in their data set, giving early insight of trends and patterns to target immediately. The system picks up proximities of terms and the frequency of their appearance within a data set, to further filter down and quickly view the documents within. Clients often discover phrases they may have missed or ones that are of no importance or relevance to an investigation.

The Concept Browser allows users to see conceptual “clusters”, i.e. themes within the data set that appear to be conceptually linked. It provides a great deal of insight in terms of how terms are linked conceptually and the thematic proximity they have in your data set, and again, trends and patterns that you may wish to explore as part of your investigation.

The Communication Hypergraph is a unique ES Locate feature that analyses email communications between all Custodians within data sets. After the parsing of all domains and individual email addresses, it depicts, in an intuitive and interactive way, the email communication traffic and patterns between individuals. By clicking and selecting just one of the lines in the hypergraph, you are able to filter down to a specific communication for analysis to see how many emails Custodian A sent to and received from Custodian B when they were sent and how often they occur. Clients can easily spot data sent to a personal account or an unknown third party.

Predictive Coding

ES Locate offers predictive coding methodology by way of continuous active learning. The machine guides clients to the relevant content in virtually any data set, allowing clients to get to the documents most likely to be relevant to their investigation. It is by far the most efficient, accurate and defensible predictive coding methodology available and the Propel team of eDiscovery experts provides clients with all relevant strategic guidance around its implementation.

End of Branch —Communication Analysis

Email Threading is part of our standard Structured Analytics functions. By limiting the review, the most inclusive emails at the end of an email thread (End of Branch), we can make reviews faster and more efficient and spot any missing emails that may not have been part of the data collection or have been deleted by users.

Smart Filters

Smart Filters are a unique and exclusive feature to ES Locate. Users can narrow down the scope of the review by leveraging over 50 different metadata fields, as basic as Custodian, Data Source and Date Sent to more advanced communication analysis such as domains, and certain email properties.

Smart Redactions

Redactions are a key element of every disclosure exercise. ES Locate has smart redaction functions which enable clients to automatically redact sensitive data such as phone numbers, SSNs and credit card numbers. It automatically finds any identifiable patterns in individual documents or across the entire document population and redacts them.

The future of data collection

The number of data sources that needs to be considered when talking about data collection is much wider due to the expansion of the “Internet of Things”. Admittedly, the bulk of the data is sourced by the different categories mentioned earlier, but with the rise of smart devices such as smart watches, AI assistants (Siri, Alexa and Google Home) as well as the wealth of different communication applications on mobile phones, there is inevitably a need to capture a lot more in Custodian and Disclosure Questionnaires and the required data collections, as all could hold potentially relevant data.

The space to watch: AI and Ethics

Artificial Intelligence has been the hottest topic in technology forums and data analysis exercises for the past few years. Humans and machines are invited to work together towards more efficient, accurate and faster analysis of data. Predictive Coding, intelligent prioritisation and Intelligent categorisation as well as predictive filtering are just a few of the features available within the realm of technology. Algorithms sit behind the science of predictive coding and lawyers, clients, even court judges reap the benefits of artificial intelligence when automation frees up valuable time for legal document review and prediction leads to the heart of a case much faster.

Technology can predict costs to be incurred, time required, human resources to be allocated, even court outcomes. The notion goes as far as virtual hearings and robot judges actually trying cases without the need for any human intervention. However, within this technological evolution lies the question of its own defensibility and how ethically valid e.g. a case tried by a robot judge versus a human one could be. Would one feel comfortable securing an end result with no emotion involved? Or would excluding all emotion actually exclude any possibility of bias? How accurate is the technology used and the complex mathematics involved when it comes to the predictions they deliver? And what happens when one wishes to appeal the applied methodology and contest the actual result?

All incredibly valid questions and ones to be addressed only through the application of technology, its trial through time and the comparison of results produced by humans versus machines. We can get closer to a more definite answer only by embracing the use of technology in the context of legal cases and overall data analysis and utilization.

For more information please contact Melina Efstathiou or Soufian Hagul.