AI Is Shaking Up the Legal Field (But Humans Are Still Vital)

Episode Notes

Transcript

Announcer:

Managing your law practice can be challenging, marketing, time management, attracting clients, and all the things besides the cases that you need to do that aren’t billable. Welcome to this edition of the Unbillable Hour, the Law Practice Advisory podcast. This is where you’ll get the information you need from expert guests and host Christopher Anderson here on Legal Talk Network.

Christopher T. Anderson:

Welcome to the Unbillable Hour. I am your host, Christopher Anderson. And today’s episode, I’m pretty excited about it. It’s about if we fit it into one of the parts of the triangle, it is production, but it’s also these days getting to be kind of everything. But the way we’re going to talk about it today is about production. It’s all the rage and rightfully so because the world literally is changing before our eyes. And the stuff we’re going to talk about today is part of that. If you’ll remember the main triangle of what it is a law firm business must do. We acquire new clients, we have to produce results that we promised them and we achieve the business and professional results for the owners. In the center of that triangle, driving it all for better or worse is you. Today the topic is, and my guest wrote the topic so we could ask him about it, what AI can’t do yet and why legal ops talent still matters.

So to me, that’s a really great topic because the rage and everybody’s talking about AI and what AI can do and as if the whole human element doesn’t matter anymore. And so we’re going to be talking about the Marvels of ai, but we’re also going to be talking about what it cannot do and why the talent is still really important to help us talk about that as our guest. Jens Erik Gould, Jens is the founder and CEO of the Amalga Group, a provider of staffing solutions for legal service providers and law firms. Over two decades of experience in staffing and recruiting financial services and communication, Jens leads a mission driven team that helps legal organizations improve efficiency through high quality operational support. He’s been featured in Forbes in Fast Company and Time. I’ve never gotten a time cover Jens prior to founding amalga. It

Jens Erik Gould:

Wasn’t a cover. It wasn’t a cover.

Christopher T. Anderson:

He held roles at Apollo Global Management and Bloomberg, and we’re going to be talking about AI and people. Jens, welcome to the show.

Jens Erik Gould:

Thanks, Chris. Thanks for having me.

Christopher T. Anderson:

Not at all. So that’s quite a resume, but the Amalga Group has been very squarely focused in people stuff, right? You’ve helped a lot of businesses, a lot of firms with their staffing, recruiting, financial services, needs, communications, but we’re here to talk about ai. So before we get into the meat of the topic, tell us a little bit about that journey. How has that journey led you to be talking about AI and what it can and more importantly, perhaps cannot do?

Jens Erik Gould:

Yeah, absolutely. Well, what we do at amalga is we build, operate, and manage teams of legal ops professionals that are located nearshore, so Latin America that work for us, legal service providers and law firms. So they basically, our teams integrate into those companies and those firms, if it’s the service provider side, it’s record retriable discovery transcription, and if it’s law firms, it’s legal assistance paralegals. We have other sides of the business as well. We’re in other industries, but we have a heavy focus on legal. Where did this come from years ago when US companies were struggling at the time with retention and with finding the best talent, we set out to fill that gap with great talent from Latin America because there’s excellent talent and it’s on the same time zone and it has greater cultural affinity with the US than say software locations. So that’s really what the value proposition is.

Christopher T. Anderson:

Well, it all makes a lot of sense. So now we’ll get into the topic of the day, which is everybody, Jens, you know this, but the listeners now, they know this too. Everybody’s talking about ai, right? And everybody’s pushing AI agents, pushing general ai, pushing generative ai, pushing agent ai, that’s the word I was looking for, ai, and pretty soon some sort of general intelligence as well or synthetic intelligence will be coming. But let’s talk about where we are today. Let’s start with what is AI really good at and offering two firms that can be helpful. And where on the flip side of that, are we starting to see or are we continuing to see, I should say, limitations in what AI can provide?

Jens Erik Gould:

Yeah, absolutely. So there’s different categories we can look at this in, but if we take the high level view, I like to say that AI is great at the what,

The pattern tasks, whereas people still are great at the why and also the should we do this, right? The why and the should we do this? So what does that mean? So what that AI is great at summarizing, reviewing, extracting, drafting. So the practical impacts of that are faster, faster first reviews, quicker tagging, faster drafting, but the why and the should we humans are still necessary and they’re still critical to make that evaluation of should we be doing this? What are the privileged concerns? What are the ethics concerns? What are the regulatory concerns? And they have to sign off on those first drafts. They have to do the final round because there’s nuance and AI still doesn’t get that nuance. And I’ll add to that, and I think we’ll get to this in some later on, but there are many legal operations functions where engineering teams haven’t yet developed a truly end-to-end solution That Handles all stages of the process, records retrieval, eDiscovery, transcription, all these things. AI is in some of the areas, but there’s no fully unsupervised litigation ready AI that runs from end to end. And so sometimes when we talk about ai, ai, yes, but it’s not doing all that yet. So we really have to sort of parse it out and see, well, what is it doing and what isn’t it doing?

Christopher T. Anderson:

Yeah, I mean, in fact, with the firms that I work with that are trying to adopt the technology and are trying to figure out where to adopt the technology, I mean it even goes to the talent of creating the right prompt. That’s the big talent today is how do you ask or instruct AI in the right way? I mean, I think lawyers who have gotten used to really working well with associates, really working well with a team, one of the great talents that you have to develop in doing that is to give great instruction, to give great direction to that team so that you get the results that you want. In a sense, AI operates in a very similar way. It takes great instruction, but you have to be great at instructing it. Yeah,

Jens Erik Gould:

Absolutely. Prompts are key. And you’re talking about prompting a chat bot, you’re talking about prompting chat, GBT, those are the basic building blocks where we’re seeing the most AI use right now and day-to-day productivity tools. And yes, prompt engineering is key for that, but I think it’s important to realize that that is where we’re seeing most of the use. It’s not in every legal process out there. We have to make distinctions.

Christopher T. Anderson:

And of course before the prompt, the ability of AI to ingest enormous amounts of data, enormous amounts of source material so that it could be really smart and responding to that prompt, that’s where I think a lot of the efficiencies are actually gained is being able to look at this huge data set that would take humans a long time. And probably at the end of the day, having in my early career, done some liver room kind of work, you can’t even keep it all in your head. As you ingest more of it, the rest leaks out the other ear. And that doesn’t really happen with the AI so much.

Jens Erik Gould:

Deep research, for example, on Chay g BT is phenomenal. I mean what it can do at scale is the next level and it’s a hugely helpful

Christopher T. Anderson:

Tool. So with all that, where are you seeing today AI implemented the most successfully?

Jens Erik Gould:

So from our law firm leader’s perspective, which I think is the majority of your audience, I think we should start there. As I just mentioned, day-to-day productivity tools are one, there’s several categories. That’s one category, drafting emails, summaries, and you’ll see the ai, we all know like Jet GBT or Grok or Claude, but you also see it in copilot and Clio and File Vine and Smokeball, all these different platforms. So that’s one. The next would be practice specific areas summarizing medical records or depositions. Depending on the area of law that you are in, there are specific tools that you can use that apply to that area. So that’s another category. A third would be contract review, reviewing large contracts, assisting with drafting clauses, comparing documents. And again, that’s where you can upload heavy number of documents and use that power to review and sift through it in a much faster way than any human could. And the fourth pillar really we’re seeing it the most is just legal research in getting insights on top of existing data. And you see this with Westlaw and Lexi and GPT research, which we already mentioned. So those are sort of the four main areas where you see it most right now in law firms. There are some additional ones. Intake pricing is starting to gain some momentum as well, but by and large, that’s where the highest use is.

Christopher T. Anderson:

That makes sense. And that’s actually where I started the show right at the very top of the show because a lot of what we’re talking about right now is inside the firm is in the production and doing that work. But you just touched on it, so I want to acknowledge it that there’s a whole nother side of it on the marketing side that’s really, really powerful where AI can also help affirm in building your marketing, in redoing your marketing, in generating new content based on old content or one fascinating one has been predicting what the market wants to hear about has been a use. So I think those are helpful as well. Let’s dive a little bit into specific solutions. One area that I see a lot of people using, you mentioned, you said Claude, but there are these tools, we’re using one right now during this recording, but there are tools like plot I think it’s called, and every online communications platform, zoom and Teams and all these have these ais that summarize, that suggest next steps, et cetera, but in transcription. So tell me a little bit about how this ability to understand language and summarize language in real time is helping law firms and where also are law firms getting a little over reliant on it?

Jens Erik Gould:

Well, in transcription, there have been huge gains. I mean, 10 years ago it was just a completely different world. Right now AI handles roughly 80% of transcription, the transcription workload, and 10 years ago maybe that would be 90% was human and error rates for what was done. Non-human were much higher than they are today. So the progress has been immense and we’ve seen much of that in the last couple of years. But that extra 20% that humans need to do is critical. AI can take the first pass or first couple passes on transcription and qa, but it’s the humans that fix the errors and apply the rules. So speech to text captures audio quickly and cheaply at great scale. It detects flag terms codes, it flags errors, but at the end of the day, it has limits, right? It struggles with accents, it struggles with jargon or slang. It can mislabel if there’s poor audio, which in depositions can be the case. It doesn’t necessarily do well with that. So you need the scope is to come in after that 80% and fix of theirs and apply context and certify the final transcript. Obviously we are not at the point where you cannot have that extra 20% that’s human.

Christopher T. Anderson:

Do you see that going away or is that always going to be a part of it?

Jens Erik Gould:

I think so, but I think we’re far way out from that. I think we will get there. Transcription I think could happen faster than some of the other areas we’ve looked at. But also there’s a regulatory component there too and you have to have sign off there and I think that’s going to be a ways out also.

Christopher T. Anderson:

Alright, well my artificial intelligence overlords are telling me it’s time to hear a word AI or not generated from our sponsors. So we’re going to take a break and let our listeners hear a couple of messages and then we’ll be back. We’re with Jens Gould, founder and CEO of Amalga group and we’ve be back in just a few seconds. We’re back with Jens Gould. We are talking about ai. Jens is the founder and CEO of Amalga Group works in staffing solutions, legal service providers for law firms who’s been doing that for a really long time. But AI has becoming a bigger factor in everything that we do and certainly is affecting Genesys businesses. We talked a little bit about what AI has been good at, where it’s being implemented the most and where law firms might be relying on it too much. So I’d like to turn our attention to records retrieval another huge, to me seems like a huge opportunity, right? Because it’s this processing of large amounts of information that could be really, really helpful. So what are you seeing about what AI can do and where the automation can’t really complete the job in records retrieval?

Jens Erik Gould:

Yeah, like this one, I see this every day because we manage teams of record retrieval specialists. AI is speeding up the paperwork, the tracking, the workflow, but at the end of the day it can’t pick up the phone and talk to a custodian. And so with records retrieval, you’ve got order entry, then you’ve got expediting the order, and then you’ve got production. What do law firms care about? What do clients care about? They want turnaround time to be as fast as possible. What really matters there is the expediting part. It’s this facility, whatever it may be, a healthcare facility or an employer has received the request for the records, but they haven’t delivered it yet. The expediting is when the record retrieval specialists are calling to follow up with these custodians, Hey, where are the records? Can you please send them? That part is very human-centric.

It’s all about, in some cases, even building a relationship. The more you build a relationship, the more you’re empathetic with these custodians, the faster you’ll get the records. They don’t necessarily always want to hear from you. So AI cannot pick up the phone and do this. So that kind of dynamics and those relationship building, that’s a hundred percent humans right now. What AI can do is it can automate records requests, it can track the systems with dashboards and reminders. It can automate alerts for pending requests, it can do basic QC checks for missing pages. But even that part, once you’ve got the records and you’re going into production, it can potentially flag, let’s say a HIPAA violation, but you’re going to need a human to look at that because you can’t mess around there. If a custodian sent you the wrong record, you can’t send that to your client. That’s a potential HIPAA violation. You need a human making that judgment call. So it’s helping with the workflow, but I think we’re a long way off to have it do everything.

Christopher T. Anderson:

Is this also the same story basically in e-discovery? Is this where the human is still needed in the e-discovery but AI is helping a lot?

Jens Erik Gould:

Yeah. E-discovery AI is doing more, I think more along the lines of what we talked about with transcription. Service providers are building or have built great AI models, proprietary models to speed up the reviewing of documents at scale and to tag and code them faster. So deduplication, near duplicate detection, email threading, all these things that the AI is doing, it’s shrinking the haystack, but at the end of the day, the human has to still make the call on what the needles are in the haystack. And so the AI can do the predictive coding or suggest the tagging, but reviewers have to go in after and they have to make those calls and sign off on the coding or do the coding and the tagging. Why? Because there’s calls about privilege, there’s reading tone and context. There’s all these exceptions that really humans need to be, that are more equipped right now than AI to handle.

Christopher T. Anderson:

Yeah, I want to push on one of those concepts though. You said it’s shrinking the haystack. And what I wanted to ask is, is it really true? And the reason I ask that is my experience with technology throughout the years has been that as we get better at being able to sift through a large amount of information and analyze a lot of information, review a lot of information, we seem to want more of it. And so as AI gets better, we want to ingest more into the machine. So is the haystack getting smaller or are we just getting a lot? Is AI helping to point out the potential needles a lot better?

Jens Erik Gould:

Well, within a closed set of documents, yes, the AI shrinks the haystack and it does it faster. If you think globally though, it’s funny. So I’ll give you an example of that. Of course, everyone wants to talk about is AI going to replace people? Is there job loss with document review? Which eDiscovery, there are a lot of cases I’ve seen where document review teams are not shrinking. In some cases they’re growing. Why is that? Well, it’s because of what you’re saying. Yes, the AI is doing more, but there’s more documents to do, right? Because once you increase your productivity, well then you realize, well, there’s more we can do. So demand goes up. And so more workflow comes in.

Christopher T. Anderson:

Even the decision about whether we could afford to do that in this case might’ve said, well, this case isn’t really worth a full document review. Now we’re like, well, we now can do that for half the cost or one fourth of the cost that we used to be able to, so maybe it is worth it. So the teams have plenty to do, at least so far

Jens Erik Gould:

Kind of tangential, but one of your previous episodes where you made a comment, I think it, it might’ve been your guest, but that you said you thought that we won’t get to the point where there’s going to be no work for us to do because humans want to strive. They just strive. Inherently, that’s what we’re born to do. I think someone made a reference to the matrix where there was a model where it was utopia and human rejected it. I agree with that, and I think that’s true in this case too. I think that we’re never just going to sit back and go, okay, well the AI is doing all the work, therefore there’s no more work to do. I mean, we’re just going to create more, I think more work to be done because that’s how we are naturally

Christopher T. Anderson:

And probably find more value in that. So yeah, I mean I think there’s a lot of room to be optimistic about this, right? Because in doing the more, we’re actually creating more value for the amount of human input, and so we each get to produce more. And I think that bodes well at the end of the day for all of us though, it is disruptive. Let me talk to you about something that you’ve written about, and I want to address this because everything we’ve talked about so far is kind of alluding to it, which is it’s not all ai. AI is not a magic bullet, it’s not a magical solution, but it is an augmentation. So you’ve talked about something called smart hybrid workflow. So where AI and humans working together to produce more or better, be more accurate, be more complete, what does that look like in practice? How are you helping AI and human operations work better together as a smart hybrid workflow?

Jens Erik Gould:

Yeah, that’s a great question. And I think the firms and the companies that really get this are the ones that are going to succeed as we go forward in this new AI enabled world. Measurement is key. First of all, you’ve got to be measuring what your AI tools are doing for you. And it’s not just speed. So yeah, you’ve got to measure speed, but you also have to measure productivity, quality, risk. That’s what makes the hybrid model sustainable. And it’s what tells you, okay, where’s the AI working for me? Where is it not and what does my team need to focus on? So you have KPIs, measurements, what’s your turnaround time to how much did this increase my speed? What’s my quality looking like? Are there error rates? Are there, is AI missing thing could have been cost, right? Are we doing it cheaper now or not risk? Are we escalating errors? Are problems being caught? Adoption rates, are people in my firm or my company actually using this and at what percentage? So doing those measurements will help you then design what the workflow should be optimally. And you’ve got to define, okay, the AI versus the human boundaries and look at what are the layers, what is the AI doing and what are the humans doing? And then what are auditors checking? And that can help you come up with a playbook for how you do this hybrid model AI human going forward.

Christopher T. Anderson:

Yeah, that’s what makes total sense. Like Peter Drucker is still well, and his words are alive and well in that what gets measured gets managed. And I guess the big question gets to be can we trust the AI to measure itself as to how well, I think the answer is probably yes, at least so far. Alright, we are talking with Gens Gould, he’s the CEO of Amalga group. When we come back gens, I want to just talk to our listeners a little bit now about what this looks like in relation to kind work you do. So when you’re looking to scale, when you’re looking to outsource in the AI era, when you’re looking at nearshore, offshore, et cetera, I want to put all the pieces together. But before we do that, we got to hear a word from our sponsors and we’ll be back. We are back with Jens Gould, founder, CEO of Amalga group, and we’ve been talking with Jens about AI and humans and how they work together and when AI is a great solution and when the human element is still necessary to finish the job, to complete the task. But so what we wanted to talk about here at this point is how to think about this in the concept of when our teams, when we need to scale our teams, when we’re looking at outsourcing our teams, we’re looking at offshoring or nearshoring our teams. How do law firm leaders decide when and where and how to do that? What should they be thinking about as they take into consideration some of the stuff we’ve been talking about?

Jens Erik Gould:

Yeah, absolutely. I mean, sort of a natural question that gets asked after all of this is, okay, so is outsourcing still valuable in the age of ai? We just talked about needing to have a human AI hybrid system. So the answer is yes, we still need humans. Okay, well, do we need outsourced humans or should it all be in-house? A lot of what we just talked about is we’ve got to have human beings checking the ai, the humans have to sign off on it. And outsourcing is a good way of doing that because at the end of the day, you don’t necessarily want to have your high salary in-house staff doing some of this verification work, which can be pretty operational in nature. And also legal work is unpredictable and there can be scaling up and down. Outsourcing offers the ability to scale up and down very much easier than in-house hiring. So my answer would be, yes, it is still valuable. Now, your original question, well, how do you think about this? There’s three main models you can outsource onshore inside the United States that gives you more control, obviously cultural alignment, client trust, but it’s the most expensive option and less scalable. You go offshore, which is far away, India, Philippines, Eastern Europe, so forth. This is the least expensive. You have huge labor pools, but the downsides here, you have huge time zone gaps and you can have more cultural and communication gaps vis-a-vis the United States.

Third option is nearshore, which means basically Latin America. So it’s outside the United States, but it’s closer. And here you have a good balance of the two. So you get the cost savings, but at the same time you have US time zones, you have better cultural understanding of the us so it gives you the balance between the two. At the end of the day, firms and companies evaluate what makes the most sense to them and what their priorities are.

Christopher T. Anderson:

Absolutely. No, I think that’s really helpful, and particularly in review, the cultural affinity, the cultural proximity is important because there’s context and both AI and people need to be better at context in order to really do a good job of review. It’s not just ones and zeros. I’ll say it one more time. Context. Okay. Alright, crystal ball time. Let’s look down the road because the one thing that’s astounding to me, and before the show, Jens and I were talking about Ray Kurtzweil and talking about the singularity is nearer and that kind of stuff. But the premise of that book, and what really feels apparent today is that not only is change happening faster, but the rate of the change of the change is happening faster. So when we think about where we’ve come in just a couple years with the AI tools and then we look forward two years, three years, and we accept that it’s not linear that it is accelerating, what do you see? What roles will AI be able to continue to take over for itself and what rules will experience legal ops talents still hold onto even down the road?

Jens Erik Gould:

I think in two to three years from now, let’s start there. AI will handle even more of the routine work than it does today. So those examples we gave of say, transcription or eDiscovery, if it’s 80 20 now in transcription, and you’re going to see that ad rise, right? Same with E-discovery and those tools, those productivity tools the law firms are using are just going to become better and better. I mean, the difference just chatt BT between now and a year ago is vast. So that’s just going to, as you say, speed up in an exponential manner. But the judgment heavy calls, the things that require client trust, the things that require ethical considerations, those will still belong to people in three years. I think they’ll still belong to people in five to seven years. And especially the nuance part, right? The context, the nuance, the things that also that require relationships with other human beings, I think is going to take a very long time to be under AI’s domain. I think that’s the last thing to go, and I think it’s up for debate when that is, but it’s going to be a long time from now.

Christopher T. Anderson:

Well, if we believe Ray, there won’t be a time because we will lose the distinction, and perhaps that’s in our future. Jens, we’ve touched on a lot of different topics here. If people want to learn more or follow what you do or what amalga does, how can they connect with you?

Jens Erik Gould:

Yeah, absolutely. Our website is amalga group.com, so that’s A-M-L-G-A group.com, just like the word amalgamate and can also email [email protected].

Christopher T. Anderson:

Excellent. Very good. Well, thank you, Jens.

Jens Erik Gould:

Yeah, thanks for having me on. It was fun to talk about this.

Christopher T. Anderson:

Absolutely. And of course, that wraps up this edition of the Unbillable Hour, so I thank all our listeners for listening as well. One more time. Our guest today has been Jens Gould, the founder and CEO of Amalga Group. And yeah, the website, amalga group.com is where you can learn more. Of course, I am Christopher T Anderson, and I look forward to seeing you next month with another great guest as we learn more about topics that help us build the law firm business that works for you. I am still a human. Don’t forget, you can also connect with us at the community table, and you never know. Jens might be there. If you’ve got questions for him, you can show up and you can ask him questions. It’s every third Thursday at 3:00 PM Eastern. That’s 12 o’clock Pacific, one Mountain two central, and in Arizona, nobody really knows because it just keeps changing. But third Thursday is at three, the community table. You can come ask your questions live, you can drop your questions right here on the show notes in the mailbag, and we will read them and answer them for you. And of course, you can subscribe to all the additions of this podcast on legal talk network.com or iTunes until we speak again, I thank you for joining us and we’ll be back. The views expressed

Announcer:

By the participants of this program are their own and do not represent the views of nor are they endorsed by Legal Talk Network, its officers, directors, employees, agents, representatives, shareholders, and subsidiaries. None of the content should be considered legal advice. As always, consult a lawyer. Thanks for listening to the Unbillable Hour, the Law Practice Advisory Podcast. Join us again for the next edition right here with Legal Talk Network.

Continue Reading