Hey, so here’s the deal, most legal teams will tell you that reviewing the same damn documents over and over is a freakin’ expensive part of eDiscovery. And honestly, they’re not wrong. With old school tools and methods, you start from scratch every single time, even if these documents are related to previous or concurrent matters.
But guess what? We got some fancy tools now that can change the game. We’re talking about artificial intelligence, concept clustering, and even traditional TAR models. These bad boys can be used in a way that takes advantage of the work you’ve already done and cuts down on those endless hours of human review.
In our previous posts, we talked about how starting small is the key to breaking free from this repeated review cycle. We also shared three ways technology can help. Now it’s time to dig deep and focus on AI, baby.
So, how the hell can modern AI help us reduce this repeated review madness?
Reuse past coding decisions instead of wasting time on more review
Now, here’s the cool part. Modern AI tools keep a record of how those documents are coded. The first time you use AI on a matter, it remembers all the coding decisions you made for each document in the corpus. And get this, when any of those documents come up in the future, the AI shows you how they were freakin’ coded before. Yeah, let that sink in. If a document was coded multiple times, AI shows the whole damn history.
Now, this is a game changer, my friends. Especially when it comes to important shit like privilege and other sensitive information. Look, relevancy may change depending on the matter, along with strategy and all that crap. But privilege and sensitive stuff, they don’t change that much. If a document is considered PII or PHI once, it better be coded like that every freakin’ time.
With AI, you don’t have to worry. Attorneys can code sensitive info consistently based on the document’s coding history. This way, you don’t risk accidentally sending that important shit to production, and you save your precious time and money by not redoing someone else’s work.
And don’t think this is all just theoretical mumbo jumbo. Firms and corporations are already using modern AI to reuse coding decisions like it’s no big deal. I mean, a freakin’ global pharmaceutical company tried this approach on 5 matters and was able to reuse a whopping 26,000 previous coding decisions. Now that’s some serious time-saving shit right there.
Beyond repeated review: get smarter and more accurate over time
Now, listen up, because this is where things get real interesting. Modern AI doesn’t just help with repeated review, it also helps you kick ass when dealing with new documents. And it does that by getting smarter as it ingests more data.
With each matter, modern AI learns and adds to its massive collection of rules that guide its analysis and recommendations. It’s like the AI is frickin’ learning on the job, understanding what counts as privilege and other classifications. It even gets all those little nuances that help define those things specifically for your data. So, the more it works, the better it gets at making those classifications.
It’s kinda like those badass review attorneys who know all the ins and outs of your data and decision making because they’ve been working with you for a long ass time. You know, the attorneys you keep hiring because they’re that freakin’ good. Well, modern AI is like having one of those attorneys, but even better. It accumulates and integrates knowledge in a way that leaves you speechless. Seriously, after using modern AI on every matter for 3 years, the eDiscovery director of some global tech company saw privilege review on a new matter shrink by almost 90%. That’s mind-blowing, my friends.
“We expected nearly 190k documents would be subject to privilege review using our typical workflow; with Lighthouse AI and outstanding outside counsel, our actual results were just over 24k.” – Director of eDiscovery, global tech company
And hey, here’s another important thing. You don’t need to keep all that past matter data around for modern AI to learn from it. Once a matter is done and dusted, AI automatically updates its rules and decision-making process. You can happily delete that matter without worry. Traditional TAR models can’t do that crap, they need the files on hand to extract and apply any learnings from them.
But hold up, not all AI platforms are created equal
Now, let me warn you about something. AI is everywhere in the eDiscovery world these days. But here’s the catch, not all AI is the same, my friends.
The inner workings vary a whole bunch. Some use machine learning which has been around since the frickin’ 70s, while others use modern innovations like deep learning and large language models. And let me tell you, the applications and benefits of different AI offerings can be wildly different.
So, if you want AI that actually helps reduce this repeated review nonsense, here are some questions to ask:
- Can this tool reuse coding? It should, dammit!
- Will the model tell me the historical coding of a document? It better, or what’s the point?
- Does it refine its model and analysis over time? It better freakin’ get smarter!
- And does it do all this without keeping data from previous matters? We don’t want that crap cluttering up our system.
And look, let’s not forget that AI is just one option for reducing repeated review. There are other alternatives that don’t even require fancy technology, just some damn forethought and strategy to make the most out of each review cycle.
If you want to dive deeper into this repeated review madness and what the hell you can do about it, check out our deep dive on the subject.