Sponsored briefing: Adoption of predictive coding for legal document review

Navigant’s Tanya Gross asks how close are we to mainstream acceptance?

Predictive coding helps tackle large document review exercises. It enables the reviewer to find key documents by training the technology to identify documents of interest (through an algorithm and selected criteria) rather than relying solely on manual human review decisions, which are often subjective. This reduces the time and legal cost associated with review exercises and gets the most relevant documents into the hands of senior counsel quickly.

Navigant has developed its own proprietary predictive coding solution, and has experience in its implementation from the more traditional model-based approach to the recent development of Continuous Active Learning. From our experience, the number of cases using this technology is rising steadily, there have been judgements that have taken place in recent years that have approved the use of predictive coding on cases [Pyrrho Investments v MWB Property; and David Brown v BCA Trading]. However, to get a sense of how this fits within the wider industry picture, I spoke with three members of the legal community: Maren Strandevold, an associate at Haynes and Boone who has worked on a number of disclosure matters both in her current position and other positions at law firms; Mark Deem, a partner at Cooley (UK) who has led teams to deliver effective results to clients in response to litigation and disputes; and Erica Albertson, a senior e-discovery project manager at Simmons & Simmons who provides support and solutions to internal practices at Simmons that have requirements to search and disclose information.

Strandevold has experienced first-hand some of the challenges involved in document-review exercises and the length of time traditional methods can take, in particular hiring review teams to conduct a first pass review exercise to identify relevant documents. Strandevold believes that the technology drives the results but those supporting you and hosting the solution need to understand how to leverage the technology to help the legal team achieve what they require.

‘In many senses, predictive coding is an extension and maturation of older technology and shouldn’t be something that the legal industry needs to fear. When used correctly the technology definitely improves searching accuracy and cuts the costs traditionally associated with e-discovery. But if you get the set-up wrong you can end up doing more harm than good. I think that’s where some of the worry within the industry comes from.’

Deem has noticed many in his profession still worry about putting their trust in an algorithm and moving away from more traditional and prescribed ways of document review. ‘There are still very tough barriers to break down,’ he explains. ‘There are a lot of people across the industry who have always done things a certain way and don’t want to change. People who are not technologically engaged and don’t see why they need to be.’

Education therefore is key and growing numbers of law firms do have in-house professionals to provide guidance on best practice. Some even have their own e-discovery teams and in-house solutions to ensure the technology is closely managed. Law firms like Simmons, for example, have had this in place for a decade and during that time have successfully built predictive coding into their workflows and trained practice teams on the benefits. Today they use it across many of their cases. Transpose this situation across the UK and we can assume there are a growing number of lawyers at all levels who work regularly with predictive coding and who have a deep understanding of how the technology will benefit their clients.

Across the wider industry other reasons surrounding lack of adoption arise at the stage in which the solutions provider is consulted and considered to assist in the process. Often consulting providers are selected from a range of different companies who offer similar solutions at the last minute under time pressure. Cost of the whole exercise is a major factor and the service and the value of the experience can frequently be an afterthought. These reactive situations can lead to bad results; cost savings are a priority for all clients but they are only beneficial if a project can be effectively managed by people who have the necessary credentials and expertise. Just like legal professionals have a biography of cases they have worked on, e-discovery professionals can share their experiences in the same manner. Sometimes, larger upfront costs will contribute to longer-term cost savings over the duration of the review process. Similarly you do not know the nature of the matter and how it will evolve until you start reviewing the information.

But do the actual law firm clients have the same concerns about predictive coding? All three people at legal practices we spoke to believe not. ‘The reaction from clients these days is actually very positive,’ comments Albertson. ‘They have heard of the technology and, in our case, because we’ve been such a long-term supporter of e-discovery, they actually come to us wanting to make use of our in-house expertise in applying that technology. Where there are concerns it comes down to two things: the comparatively high up-front costs, which we can explain to them ultimately result in lower overall project fees; and concerns that not every document is going to be read by a human.’

‘We have run enough predictive coding projects to demonstrate to clients that they will ultimately save between 20% and 50% on their document review fees,’ she continues. ‘And we can assure them that it’s simply another proven tool in our e-discovery arsenal.’

Strandevold concurs: ‘For us, the real concern that we see from clients relates to handing over a wide range of data. Once those concerns are dealt with, the mechanics of the analysis doesn’t really concern them that much. We’re the experts and they trust us. Basically, our clients want the whole e-disclosure process to be efficient and, as predictive coding can increase efficiency and keep costs down, they generally welcome its use on their case if we advise them that it is appropriate in the circumstances.’

For Deem, it also comes down to how comfortable lawyers themselves are with predictive coding. ‘If lawyers are confident in their explanation of predictive coding to clients, then clients are usually happy to go down that route.’

In conclusion, from talking with legal teams that regularly use predictive coding in their document reviews it is certainly something that will continue to be used on their cases, and its prevalence will only increase. ‘Predictive coding has the potential to revolutionise the document review process,’ concludes Deem. ‘It lets legal teams create an early assessment of the documents and decide much sooner where the case strategy should be based.’

Albertson concurs: ‘This really is the most important benefit for us. We get to the most important documents really quickly and start to form an idea of how to conduct the case much earlier on.’

Data volumes show no sign of decreasing and predictive coding offers a way to keep document-review fees in check even against the rapidly growing amount of data that is required to be disclosed in many of today’s cases, especially as the technology continues to evolve. Making sure the team you are working with has the experience and the credentials to support the workflow is also a key success factor.

Sponsored briefing in association with Navigant