Artificial Intelligence (AI) thinks, learns and problem solves more efficiently than humans. AI is all around us and in almost everything we touch, it is an algorithm that is designed to make our lives easier and is sometimes referred to as machine learning.
In the case of litigation, it can save time and money by streamlining the process of document review, eDiscovery, and preparation for forensic cases. Computer Forensic Expert, Lee Neubecker and Catherine “Cat” Casey who is the Chief Innovation Officer for DISCO discuss how AI works to improve legal technology.
DISCO is a leader in legal technology is a developer of a cloud-native eDiscovery software for law firms designed to automate and simplify error-prone tasks. They provide a myriad of different types of analytics that will supercharge searching data dramatically reducing time and money.
Part 1 of our Three-Part Series on Artificial Intelligence (AI)
The Video Transcript Follows.
Lee Neubecker (LN): Hi, I’m here today with Cat Casey from CS DISCO. Thanks for being on the show.
Cat Casey (CC): My pleasure.
LN: We’re going to talk a little about artificial intelligence as it relates to eDiscovery and document review. Cat, can you tell us just a little bit about what your firm does to help speed up the review process and lower costs for clients.
CC: Absolutely, we’re a cloud-native AI-powered eDiscovery company. And what that means is we’ve got vast amounts of elastic computational power that we can use to run a myriad of different types of analytics on data to supercharge your searching and dramatically reduce the amount of time it takes you to get to that key actionable evidence. So, we’ve kind of flipped everything on its head. Instead of being a question of how quickly can I read through all of this data, it’s how laparoscopically can I surgically find all of that key information. The results that we’re seeing are pretty resounding. Up to 60% reduction in time to get to that key evidence. Freeing up attorneys to get back to what they went to school for, the practice of law. It’s pretty compelling. We’ve had some pretty interesting additions, including even today, we just announced, I think, the first true AI in eDiscovery with AI model sharing. Basically, with each iteration, with each type of case that you conduct with DISCO, our algorithms are getting smarter. We’re extracting insights and building in more robust taxonomy and analytic structure to parse data, which is going to yield better and better results for our clients. It’s truly exciting.
LN: So we’ve come a long way from the early days when the attorneys wanted everything printed and Bates-labeled before they looked at it. To now, moving ahead using TAR, technology-assisted review, like artificial intelligence, which fits into that, correct?
CC: 100%, we have a continual active learning model, so it’s more reinforcement learning than a standard supervised learning model. Basically, from the coding of document one, our algorithm’s getting smarter and making recommendations on highly likely to be similar documents. We battle test the algorithm on an ongoing basis. Whether it is an affirmative or a negative for a suggested document, the algorithm learns more, and because of that, we prioritize the most relevant information quickly and people are able to then accelerate their review speeds by up to, I think we’ve had over 180 docs per hour. So, it’s pretty compelling and this is just the beginning.
LN: So your platform’s all in the cloud, correct? So companies or law firms, they need no infrastructure other than a browser?
CC: 100%, the nice thing, in my prior life, I ran a global discovery program, and I spent hundreds of thousands of dollars a year just to keep pace, just to have storage, just to have basic replication and back up, and all of that. Now, even a small firm, all the way up to an Am Law One firm or a massive Fortune One company, they can have the same robust technology without having to set up a data center, without having to invest a ton of money. It lets everyone level up and has a better experience throughout the discovery process.
LN: One of the challenges a lot of my clients always have is they have a need to understand what the costs are going to be and to be able to communicate to their clients those expectations so they’re not throwing their clients on the eDiscovery rollercoaster of non-controllable bills. How does DISCO help to address those concerns?
CC: Transparency is a major pain point. One of the banes of my existence used to be trying to normalize this pricing model versus this, versus this service provider, versus this technology. We just throw that all out. We charge one flat amount per gig. It includes analytics. It includes processing. It includes everything, and we work with you to get the volume of data that is being applied to that one flat cost per gig down. It eliminates that hide the ball gotcha moment and it gives a lot of transparency. And of course, if someone wants a different model, we’re happy to accommodate that. But in general, straight, simple, honest. It’s really rewarding for our clients.
LN: So, what cases, what types of litigation case matters do you see as having some of the best benefits of being migrated into your platform?
CC: Yeah, I think any case can. If you’re a tiny company, it helps you be David versus Goliath. Even on a small data volume case, you can start getting insights and reduce the amount of time you’re having to spend doing something maybe you can’t chargeback for. For a big massive case, because we are an AWS and we were built on kind of convolutional neural networking, we’re moving, and we have such a robust computational lift, even we’ve had 150 million documents with hundreds of users and we still have sub one second page to page. We are still lightning fast. And so, whether it’s a big case, a simple case, a complex case, there is a value proposition for almost anyone.
LN: In terms of the types of law firms that are using your platform, do you see many smaller, medium-size firms using your–
CC: Tons, actually tons. That was where we got our teeth. Boutique, we started as a boutique law firm. We actually were a bunch of attorneys that were frustrated that all the tools were terrible, and so they built their own. And so, the foundation of DISCO, we had a family of tons of boutique law firms that we were supporting, we still do to this day. The tool we built though, had a longer vision. It was built to be much bigger and more scalable, and as a result, that’s why you’re seeing us with major, the WilmerHales of the world, very large firms and very large corporations because the tool itself can scale up so much.
LN: Great, what are some of the challenges of working, that law firms find that already have entrenched solutions? There are other review products out there and if they really want to make the benefit of your platform, don’t they have to kind of fully use it for the case?
CC: I would say you probably don’t want to split the baby with a case. If you’re processing with another tool, you’re not going to get the same benefit as working with DISCO. But you don’t have to move your entire litigation portfolio to DISCO day one. We’re seeing a lot of people that are sunsetting Legacy Product and Legacy Platforms moving towards DISCO, but it’s not, “I’m going to move every single case today.” It’s going forward, we’re going to start bringing in new cases. There tends to be such an improved experience and improved UI for the attorneys that they start to not want to use the other technology as much.
LN: I know as a computer forensic expert, oftentimes we’re going out initially collecting and forensically preserving the data. But your product sounds like it would be right for a firm that does forensics that needs to collect different data from computers, possibly harvest just an email. Filter the dates and times of the email to a PST and then they can take those PSTs and upload it into your platform, correct?
CC: 100% and we also, we’ve productized some advanced ECA, where we charge a much, much lower rate. So, you get three months no cost hosting. It’s half the usual rate, and you can do ECA for up to three months. And the goal of that is to let’s whittle down to the most surgical, teeny, tiny, laparoscopic piece of data set that you can have. An example was we had a 20 million document case and we were able to run the ECA, get it down to about 5.6 million documents. Run more coaling, run our analytics, get it down to about 200,000 documents. And usually, that would be when you have to review every single one, but we were able to, with our workflow, with CAL, get it down to 140,000 documents. And so, if you think 50 bucks an hour, an attorney can only do 50 docs an hour, the cost savings is monumental.
LN: So as someone uses your platform and they start to tag and prioritize certain documents, your software learns based on that taking. It helps find related concepts to those conversations and what not?
CC: 100%, 100%.
LN: So really, the more that are reviewed as responsive, similar concepts and whatnot so that important links aren’t missed.
CC: 100% and because we do automatic batching, is every new batch of documents a person gets because we’ve applied this artificial intelligence and continual active learning model, it is a more relevant subset of data and people are able to go through it more faster. And sometimes, they will get to a point where they can say, “I’ve hit all my relevant information. “The rest is not relevant. “I’m going to sample it and statistically determine “I don’t have to review those last 100,000 documents “that maybe aren’t relevant,” and it’s pretty cool.
LN: In our next segment, we’re going to be talking What the trends are in the industry impacting law and eDiscovery. And then finally, we’ll talk about some of the pitfalls of what companies, organizations, and law firms face if they don’t embrace artificial intelligence to help make their review process more efficient. Well, thanks for being on the show.
CC: My pleasure.