Luke and I discussed differences between civil law systems, such as they have in Belgium, and common law systems such as we use in the US; the imperative for greater efficiency driven by the constraints of civil law systems; how eDiscovery in Europe is driven more by investigations than by litigation, with Luke pointing to merger clearances as an example. We turned to the slow but growing adoption in Europe of predictive coding, advanced analytics, and AI. I had Luke gaze into his crystal ball and tell me what that adoption might look like in the future. Finally, I asked Luke what his dream eDiscovery platform would look like and he came back with the most expansive vision yet.
Recorded live on December 11, 2020 | Transcription below
Note: This content has been edited and condensed for clarity.
Welcome to eDiscovery Leaders Live, hosted by ACEDS, and sponsored by Reveal. I am George Socha, Senior Vice President of Brand Awareness at Reveal. Each Friday morning at 11 am Eastern, I host an episode of eDiscovery Leaders Live where I get a chance to chat with eDiscovery luminaries.
This week, I am pleased to have joining us, Luke Smith. Luke is the founder, CEO and Head of Customer Delivery at LTI – Legal Tech Innovations – in Belgium. In terms of his experience, Luke’s been in the industry for a while. Prior to founding LTI, he was at Freshfields as Head of Antitrust Investigations in Belgium; before that, at Latham & Watkins; and before that, at Epiq. Prior to that, he was with Hogan Lovells in London and Linklaters before that. Luke has had a long and extensive tenure in the industry with experience both on the provider side and the law firm side. Luke, welcome.
Hi George, thank you for having me.
Civil Law and Common Law Differences
Pleased to have you here. So, let’s get going. One of the things I’d like to focus on, you are the first of our guests who works in a civil law system, as opposed to a common law system, which we have in the United States or they have in the UK or Ireland. There are all sorts of implications when it comes to eDiscovery. Why don’t we dive right in and start talking about some of the differences and what those mean.
Where it came from originally was Linklaters and Hogan Lovells in the United Kingdom, and I started working in the eDiscovery document management world for cases in 2000 – about 20 years ago, almost 21. I moved to Belgium around 11 years ago and one of the things that surprised me when I moved to Continent, continental Europe, is the lack of discovery and disclosure that you would typically find in the UK and US systems. It took me back a little bit because in the United Kingdom we saw large volumes when there was a litigation or a dispute that needed to collect such, review and then more importantly produce these vast volumes of documents from your system into the other party’s system and visa versa.
In continental Europe, things play out a little bit different for lawyers. Here in continental Europe, you do have some large discovery and disclosure of evidence exercises but they’re very rare, they don’t happen so often. One thing that took me back a little bit was the need still to have care and attention when presenting your evidence, because here it be more of a trial by surprise. The lawyers would be liaising with their clients to talk about what the dispute is, ask them for evidence, and then they would be able to build their file and their bundle of documents that they would use in a hearing. There wouldn’t necessarily be a need to actually hand over everything in the universe that relates to that, it would just be the core documents that you’re focusing on.
There’s a need still, there is a practice to handle this document and the evidence with care. There’s more of a forensic background experience in continental Europe versus an eDiscovery background. You would engage with a digital forensic expert and say: okay we have this person, we have this case, we now need you to help us find the evidence, collect it in a way so it can be authenticated in case the judge or the other law firm’s lawyers ask to prove its authenticity. The knowledge of eDiscovery is less profound in continental Europe. A lot of the time when I first came across to Europe, I’d be referred to as “eDiscovery” or “eDisclosure” and the lawyers would be, “I don’t do U.S cases so thank you but I don’t need your help.”
“I think the key differences in all the disclosure rules, discovery rules, is the way you exchange these vast volumes of information is a bit different in the litigation and disputes worlds.”
A lot of the time the terminology is called a forensics search or an eSearch, but in principle it is the same thing as eDiscovery, because you have to find your evidence and your facts. On cases where you have lawyers working with their clients who aren’t able to pinpoint the evidence to you, you have to go through an exercise of collecting a wider set of documents and trying to build your case: go through the documents, find the facts and build your core bundle. Again, I think the key differences in all the disclosure rules, discovery rules, is the way you exchange these vast volumes of information is a bit different in the litigation and disputes worlds.
Similar Needs, Different Latitudes, and an Efficiency Imperative
Whether you’re in a civil law system or are a common law system, ultimately you do need to present a story, you need to support that story with evidence, which means you need to figure out what story you’re going to tell and what evidence you can use to support it. We have a very wide reaching process here in the US, where attorneys can range far and wide, looking for data, pulling a lot of the data – they have a lot of latitude in terms of what they go after – and also spent a lot of money on all of this. It sounds as if in Belgium, for example, the need is the same – you still need to tell a story, it needs to be a persuasive story, it needs to be supported by evidence – but the latitude is very different. You have to be much more disciplined and focused in how you find, get and go through data and then you only hand over a very small portion to the other side. Am I understanding correctly?
That’s exactly right. Every now and again, you do get such a large dispute or litigation where you do have the need to cast a net very wide to find your evidence, the same as you would in the US or the UK, but these cases are much less frequent. The size of the industry and then the actual knowledge of conducting these exercises in the legal world is much, much smaller. But the need is the same.
With the need being the same but the latitude being much different, how do folks efficiently and effectively accomplish all of this? We know how we do it in the United States; clearly you have to do it a different way. How are you approaching it?
I’ve had over the years the opportunity to work with many, many, great lawyers and one of them was a DC competition lawyer who came across to Europe and he coined the phrase, “the Wild West”. He said, “You just have no rules, you just do what you believe is right”. And he said it’s like the Wild West.
It takes a lot of care and attention because as you appreciate the knowledge of conducting these exercises is much smaller, and this also hurts, perhaps, the financial aspects of what a client may pay. If you are forced into conducting such a large eDiscovery exercise and you have dozens or tens, hundreds of custodians to investigate and collect data from and process hundreds of terabytes of data, you could be looking at a tens or hundreds of thousands of euros or dollars eDiscovery bill before the first search term or before the first page is even turned in the review. This can sometimes make it hard from an efficiency standpoint, because before a budget can even be set you may have backed into a lot of cost here. You have proportionality cards, I think that also exists in the US and in the UK too.
I would always recommend to make things efficient. Start with your key actors, the key custodians. You pivot. You process some of the data. You start your investigation, your research into the internal documents, the evidence, at a very small scale. You don’t throw hundreds of paralegals, hundreds of thousands of documents. You start into bite sized chunks and then pivot based on what you’ve learned. If you’ve got lots of irrelevant documents, why are these coming back? You change your approach, pivot again, and run a new set of searches or try at some analytics and see what that yields. If the responsive numbers are greater, then that’s a good process; we keep doing that until perhaps the responsiveness trails off a little bit. From an efficiency standpoint, it’s definitely start small and then from your learnings, bring more custodians in, more actors into the investigation or the review as you progress.
The Driver in Europe: Investigations, Not Litigation
eDiscovery is a portion of the activity that goes on for litigation, what we would call civil litigation, a portion of the activity that goes on in Europe, but it’s by no means all of it, right? There are investigations of all matter?
“The regulatory investigation work and corporate investigation work has really driven the adoption of eDiscovery in Europe.”
Talk about that a little.
Since I’ve moved to Europe, my number of litigations I can count on both hands, but the number of investigations now is, over the last 10 years, probably over a 100 investigations. The regulatory investigation work and corporate investigation work has really driven the adoption of eDiscovery in Europe.
Again, these are not as frequent as a dispute may be, but they are definitely rolling along now. All of the Bundeskartellamt in Germany, the ACM, the Dutch, French, the Belgium competition authorities and the European Commission have become a lot more wiser and knowledgeable on these larger exercises of uncovering evidence. These regulatory investigations are definitely driving the eDiscovery industry in continental Europe.
One area that stands out to me is the merger clearance. When you have two dominant companies merging in Europe, a bit like the DOJ process, the Phase 2s, of course the European Commission will act on this merger and they need to approve it, they need to do a market assessment: How is this going to affect the power of this company after the merger, our suppliers, consumers, people like us? Will prices go up? They have now used their powers to request large numbers of internal documents from the merging parties and the law firms are responsible for taking care of this – collecting all of the evidence, the internal documents and providing it to the regulator.
In Europe, they issue something called a “Request for Information” or an “RFI” for shorts. These RFIs are 12 or so paged documents. They ask for the documents to be handed over that relate to the planned investments that the companies have had: actual planned investments, the innovation, products that they’re looking at implementing, how this is going to affect pricing from suppliers, customers, how it will affect certain areas within the European Union. Very often these transactions are very big, so they can also be very political. If you’ve got a large US company buying out a European company, maybe the President of France is not so happy about this and wants the Commission to be very thorough.
They issue these RFIs with say 20, 10, 30, 20 different topics of interest that they want the law firms to look for, to search for, and then hand over anything that’s responsive. These RFIs normally contain a list of the employees or custodians they may want you to search. This will be the VIPs, the CEO, the head of sales, head of marketing, lots of different products and business areas. They also target now people lower down the chain to see how they may be discussing it differently to the people at the very top. Part of these RFIs now, they’ve changed a lot in the last few years, this is where it gets our attention, so the eDiscovery professionals’ attention is: they ask for the collection logs. When you’ve imaged the phones, the laptops, collected data from the e-mail systems, servers, you have to provide the logs of the collection over.
When you are crafting your search terms to answer the 20 questions, the topics that they’re after, you have to present your search terms that you are using to find these documents. Then you have to provide the load file – which for a lawyer who has not done any eDiscovery work, it takes you back a little bit – with the classic metadata I think most people would be used to providing on this call; they ask for the path of where the document originated from in the environment. Gone are the days where the lawyers could just say to the custodians, email me everything and I’ll copy it onto a USB stick and hand it over. We actually have to do this exercise in an eDiscovery-style fashion.
The regulators are definitely responsible for enhancing the knowledge and the need and the adoption process of eDiscovery workflows, best practices and using the technology efficiently.
Predictive Coding and Advanced Analytics: Slow But Growing Acceptance and Use
Are the regulators looking for certain types of approaches when it comes to not just gathering the data but evaluating it in order to determine what to hand over? Are they looking for certain types of technology to be used, or do they want certain types of technology not to be used?
Many of the regulators now have lawyers who have worked at law firms, who have moved into the regulatory body. They are perhaps familiar with the technology that can be used to search and present the results of search terms so they can analyze them and see if the terms are producing too many hits or not.
Very recently, over the last year or two, there has been some debate about whether you could use predictive coding or analytics to help uncover the documents and this will be used in tandem with search terms. It’s something I’m in favor of as a technology lover, but it’s somewhat quite challenging in Europe to implement because normally these exercises need to be completed in three to four weeks, you can get a little bit of an extension. They’re quite hard because you’ve got 20 very different topics and then you may have Spanish, Italian, French, German and English languages in the data. It’s quite hard to recommend to lawyers, “We can run this predictive coding model for topic number 2”, because we would have to think, “Okay now, we’ve got to run in different languages, do we have enough time, do we have enough resource while we’re looking at these other 19 topics in tandem?” and you’ve got your privilege reviews going on. It makes it a little bit hard to implement the more sophisticated approaches to find any evidence, just purely from the timeline base.
Is it lack of familiarity on the part of the lawyers that’s the key barrier here?
I would say it’s more in the regulators hands, the lawyers on the regulator side. They hold the cards to some extent because now they dictate a little bit about what search terms you will apply to which topic and then you can have a discussion about amending them if they’ve happened to use the company name as a search term, which sometimes happens – returns every document.
I guess it is an appreciation of the technology, what it can do and how it can be achieved within the time that you have. Also, the regulator has the clock running; they only have a certain number of days to actually approve the transaction. If you get bogged down in a four-week discussion about the scores – the results of the predictive coding model – and then spending time to actually qualify the results, they lose too much time themselves. Because you’ve got to then hand over these 500,000 documents and then they’re going to have to run their searches and conduct their reviews as well within that 30-odd day time period. It’s not a lack of appreciation, it’s that the time is of the essence, so it’s quite hard to agree on that when it’s already difficult to agree on search terms, which you would think is a little bit more easy.
Gazing Into a Crystal Ball: Advanced Technologies and AI Moving Forward
Sometimes, sometimes not so much. I have got one more question in this area and then I want to move on to a different topic. The question is, go ahead and take a look at your crystal ball. If today there is reticence on the part of regulators to use those more advanced sorts of technologies, protective coding and other forms of artificial intelligence, what do you think it will look like one, two, three, five years in the future?
I would hope, I would imagine, that the regulators will be conducting their internal reviews of the data that they’ve been receiving from the parties who are merging or they’re investigating, even if it’s a cartel. I would predict they should now be banking the knowledge and creating these models, their own internal models, about, “This is what innovation looks like, the topics of how we’re going to enhance products or services into the market”, or “This is what a price war or cartel or customer-sharing, bid-rigging information looks like, these are examples that we have”. Crystal ball, maybe in five years, we may actually see the law firm say, “Okay we have a database with 3,000,000 documents”. And perhaps the regulator would say, “Okay we want you to run this model, which is bundled up of text of previous examples of words”. And then ask them to assign it into their dataset. Then they would need to provide the top 1000 documents.
I think the market is somewhat there now. We have certain tools. They may not be compatible with the different products on the market, but if we could have a standard way of implementing these models, then that could be done. I think the regulator will start having this historical bank of information that will be turned into AI models that they will then implement on the datasets.
The Dream Platform: Augmented Reality Analysis
So now, a different crystal ball. I’ve asked this of a couple of other guests before. If you could have your dream package or dream platform, your dream technology capabilities for use with eDiscovery, for litigation and investigations, what would that look like?
“I would love to have the datasets presented in an augmented reality way. You could have these shapes and circles representing the data. It would be voice enabled, augmented reality.”
A great question. If it’s a dream and anything’s possible,I would love to have the datasets presented in an augmented reality way. You could have these shapes and circles representing the data. It would be voice enabled, augmented reality.
You could say, “Show me in a dataset where the competition is mentioned, the names of competitors in the market”. They would come up and you would be able to drag them over into a bucket, and then you would say, “In that bucket, show me who the key actors are”. It would come up with pictures, faces. Then you could pick which ones you want, you could drag them into another bucket and then within that bucket, you would then say, “Show me the people who have been active in an aggressive nature”. You’ve got these filters going on with the buckets, and then the documents will come up and then you would say, “Right, we’re going to put them into a review batch”. Then it goes into a typical style document review exercise. That would be the dream, to have a platform you could talk to, you could interact with, and really cut up and carve the data.
I’ve got this picture of you standing there in front of a screen that doesn’t even exist, because you’ve got the Google glasses or what glass or whatever it is. You are moving your hands around. To an outside observer, they can’t figure out what you’re doing but for you, you’ve got this three-dimensional universe of data in a structured fashion in front of you, where you are able to grab pieces, expand them out, compress them down, move them around, combine and recombine. That sounds like what you’re talking about.
Yeah, like the movie with Tom Cruise, which was probably almost 20 years ago, I guess now. Something crazy like that would be….
Well, it might be crazy, but it would be really cool, wouldn’t it?
Yeah. Well, yeah, yeah. Some of the technology is there, we do have some now that’s able to detect stress and passive and aggressive tones in text. So it’s just trying to package that up in a sort of augmented reality way. That would be my dream if… Hopefully we may even see that one day.
Well, I will conclude that part with the same thing I’ve said to others, “Let’s see if we can make that happen”, because that would be a lot of fun. Thanks Luke. Luke Smith is with – do you go by LTI or Legal Tech Innovations? Which do you prefer?
Yeah, Legal Tech Innovations is a bit long, so I just cut it down to LTI.
Okay. Luke is Founder, CEO and Head of Customer Delivery at LTI. Luke, thank you very much for joining us this week.
Next week we will have with us Joy Murao, who is Founder and CEO of Practice Aligned Resources.
Luke, once again, thank you very much.
Thank you George. Thank you very much.