AI’s Impact on the Legal Industry – Techstrong AI Podcast EP63

Hello and welcome to the Techstrong AI Podcast. I'm Amanda Ani, and with me today I have Tom Dunlop. He is the CEO of semis.

How are you doing today? I'm great, I'm great. Thanks Amanda.

Thanks for having me. Can you share a little bit about semi what services do you provide? Of course.

Yeah, so IZE is a software as a service, uh, solution for legal teams. Um, and essentially we want to make every interaction with a contract more efficient, so we're an end-to-end CLM. Um, and really what we focused in on as a, I guess a key differentiator is to kind of embed our UI and our experience into the tools that people use every day.

So think Teams and Slack and Outlook and Gmail. Um, and really what we're trying to do has bring the power of ACL M, but embedded within the tools that those kind of, uh, corporates will use, uh, every day. Okay.

And now semis recently released a report, it was about, um, AI use in legal teams. C can you share a little bit about the results of that report? Yeah, of course.

I mean, I think the, you know, the legal services report that we've done, um, for, for a few years now is, is a real kind of, um, it's a, it's a real kind of useful bit of information updates on, on how legal teams are feeling their role, um, and just how we can see shifts in the market over time. And I think when you're an in-house lawyer, um, which I was before founding, uh, surmise, um, trying to find the insights from the wider market is, is really useful. 'cause sometimes it can be quite a lonely place being, you know, within a legal team, within a wider, um, a wider organization.

So I think with this year's report, um, you know, we were, we were keen to really understand the impact of ai. It's obviously, you know, huge for every corporate and every probably every single, um, area of the business as well. Um, but particularly for legal teams, I think what we were very keen to understand was, you know, they're, they're in this kind of slightly awkward place where they have to be the voices of risk and kind of governance within an organization, which clearly from an AI perspective, there is a lot of noise about the risks and the privacy concerns.

Um, but similar there that, you know, the actual departments themselves about how they work is, is kind of ripe for disruption or use of ai. So they've kind of got this kind of, you know, they need to be able the adopters of the tech, but they also need to be the kind of barrier to the wider business potentially around any kind of governance and risk concerns. So we were just really interested to understand, you know, how, how are people feeling about ai?

Has their role changed in the wake of this kind of huge shift in the market? And I think the, the general overview of what we found from the reports is, is kind of validation of that really, that, you know, I think one of the biggest stats was that three out of the four, um, legal professionals set a role have evolved over the past two years, which is quite a, quite a big change. I mean, that's, um, you know, there's not an insignificant kind of shift in the market.

And I think there was kind of a general consensus that the, the reason it's changed is the, the kind of, well, one, the macroeconomic environment and the kind of wider, um, market. So compliance and risk have really come to the forefront just with not only ai but also, um, I guess the, the wider kind of economic changes that, that are going on. So it was, it was kind of a good validation point.

Um, and I think that it kind of confirmed what we thought, but I think one of the other positives that we got out of the report was, um, that the actual adoption and appetite of AI within legal was actually really high as well. And there was, there was quite a number of people already using AI in kind of day-to-day life. So that, that was good to see.

'cause I think there's probably a perception that legal teams are relatively slow to adopt tech and ai, but that, that was kind of a good validation point as well. Let's talk about some of the top areas of concerns that they shared. I was looking at the stats and it said privacy and security, 45% limited understanding or training, 37% lack of clear use cases of value, 31%.

So let's talk about those, um, that's pretty significant. Uh, how can legal teams address these areas of concerns? Let, let's start with, I know privacy and security, that is a big one, so let's start there.

Yeah, I think what we've found with, I mean this, this generally happens in, in a number of areas where almost the consumer use of AI with exposure to chat GBT and Claude and a number of these other models kind of overtakes the enterprise adoption of these tools. And so I think what we've found over the past couple of years is, you know, individuals in their personal life are kind of experimenting with chat GBT, understanding how they can use it. Um, and then that started to creep into the workplace.

And obviously a number of these tools, um, are essentially public training models. Um, so legal are the first ones to probably make the connection of, well, hang on a minute. What, what information are you actually putting in there?

Are you trying to almost use it for your own personal capacity, but actually to do your work? So are you trying to put information in there to, I don't know, whether it be drafting emails, whether it be, um, redrafting articles or internal memos that, you know, things like that which actually could have some pretty significant confidential information. Um, and I think that's where, you know, the, the enterprise has struggled to keep up because you need to then formalize, well, what is our policy as a, as a business?

Do we have a formal what one tool that we're allowed to use internally? Um, and are we okay with rolling that out? Is there a cost?

What can you put in there? What can't you put in there? And a lot of this information was not necessarily being defined yet the usage and adoption of these products was growing pretty significantly.

So I think there was just, you know, particularly for the first year, 18 months of the kind of this really mainstream kind of AI, really in the, in the personal, um, capacity, really, it, it was a case of legal teams trying to, trying to catch up. And I think the, the enterprise looked at the legal team for guidance on this to say, you know, what, what should we do? Is there a particular model that's good?

There's not like what you tell us and you advise us what information we should put into these kind of models and what, what we shouldn't. And, and legal then had a steep learning curve. So I think part of the, you know, that the stats that have come out is, is it's kind of showing that they're, they're, they're kind of balancing two things, which is their understanding of the AI itself and the privacy concerns, and then how they're then actually applying that and advising the, the wider organization.

Um, so it's been, you know, you can, you can see that across a lot of the responses and I think with, with our customers as well, that there's this overwhelming kind of, um, steep learning curve as well as, um, kind of very quickly being asked to, to create some pretty significant policies across, across the business as well. And I think, I think that's where those stats are, are kind of, uh, reflecting that From your experience, talking with business leaders, are there any companies that have taken initiative to give a better understanding or provide some kind of training? Um, do you have some some tips in that area?

Yeah, I think a lot of the company, obviously when, um, chat PC came out, there was the kind of copilot, um, tool, which was one of the first ones that was more of an enterprise wide rollout. And I think a number of a, a few of our clients tried to roll that out as a baseline. Um, and I think what I've found where the best adoption has happened is they wrote may roll out something like a, um, a, a copilot and just say, look, this is our almost like generic enterprise tool.

It's great for querying our SharePoint or our intranet or so something in, in terms of their internal, um, database. However, they've also adopted a policy of realizing that almost like vertical specific or, or department specific AI is necessary. And I think we've had that kind of, or there has been in some cases that friction where copilot can do everything.

And I think there's the now realization that, you know, actually you do need specific AI tools for specific roles and obviously legal is, is part of that. And there's a number of other areas that, that are part of that. So the best adoption I've seen is, you know, set the baseline with a, a, a tool that can be rolled out for your email redrafting for your, you know, general research purposes for your assistance with a first draft of a, you know, a PowerPoint, for example.

Like great tools to, to really get that kind of, um, I guess first draft done and the kind of generic tool that you could use every day. But also they got quick to establish, well, where and how can we roll out specific AI tools to make individual departments also, um, you know, benefit from that so that, that they, they've been the most successful rollouts that I've seen. And, and obviously that, you know, is generally where we come in 'cause we're an AI tool for legal team.

So, um, that's where we've seen, you know, the best adoption from, from legal as well. So what are some common use cases for AI in legal work? Can you share those?

And then where are some use cases that are being overlooked? What are those and um, what ideas do you have for implementing those? Yeah, and it's been an interesting learning journey I think because when, I mean this, again, it follows that kind of the consumerization of, of ai, which was a lot of people used AI for very simple tasks.

And so I think the first use case of legal found was, um, and this is what we, we saw like a simple redraft of a clause, you know, could do you have a clause in a contract, it's quite a simple block of text, can you redraft this to be mutual or, um, could you redraft and make this, you know, more friendly to the supplier? Those kind of things. And it was almost quite simple tasks that were really reflected probably how they interacted with, you know, chat two BT for example.

Um, and those, those kinda went out, it was great. They were kind of quick efficiencies and it was very quick to get up to speed. So that was, that was great.

There's a great option curve I think where, I mean, where we see the opportunity and what we are doing with, um, I guess now as the, the AI has evolved to be a bit more agentic and how you can be a bit more complex in your, in your workflows, we can tackle more complex automation. So for example, not only would you review, let say an entire contract in one go, but one of our tools that we roll out actually does kind of a three step process. It, it kind of finds everything that's relevant, it then does an automated red line, it can then actually create tasks on the back of that and actually workflow and all of those are kind of completely automated.

So you're actually using several to do, um, a kind of an automotive workflow as well as actually taking action or suggesting action and not just this kind of one way question and answer type use of the ai. So I think what we'll find is just, you know, over the next 6, 12, 24 months as this kind of technology gets more, um, widely adopted is, is is really the, it is the complexity of task that AI can handle. And what legal teams will be able to do is just kind of be the orchestrator, just set, set the boundaries, set the parameters of the workflows and where they want things to trigger, but actually let the AI do more of those workflows and more of the heavy lifting so they can just sit there and be a bit more advisory and kind of, I guess, um, you know, focus on the more strategic strategic work, which is the goal of every in-house lawyer and legal team.

So AI is advancing quite rapidly. What do you envision for the future of legal work, say a couple of years down the road as it relates to ai? So I think it, it is kind of a, a continuation of what I was, um, talking about.

I think the first step for me is, um, almost a consensus that certain tasks are no longer done by, you know, legal teams, whether it be our legals or junior lawyers or senior lawyers. Um, and there's just the general acceptance, the, the wider business can be a bit more self-sufficient. Um, 'cause there is still a little bit of resistance there.

And I think that what I, what I expect to see is AI just, just enabled the wider business to create the first drafts of documents to maybe do the first pass review, um, of when a, you know, a, a document comes about redline like with guidance from, um, you know, a a some kind of, um, assistant in the, uh, in, in Microsoft Word for example. So I think there'll just be a general consensus that these are just how lawyers will work and they will not get involved anymore on basic drafting and those kind of what I class as red flag reviews of low value agreements. So I'm talking very specific on contract side of it, but I also think the other thing that's quite interesting move in the ai, uh, era and what we'll see in legal teams is almost the productization of knowledge.

And what we find is we have things like playbooks and we have this really specialized knowledge that's been built up over years and lawyers are, um, you know, quite precious about their own ip. They, you know, this is particularly prevalent in in law firms, but I think even in, in in-house you kind of, you, you know how to review contracts and you're overlooked in almost to write that down and share that knowledge or it just takes too much time. So I think the next big wave, what we'll see is not just the automation of tasks, but how can you actually productize knowledge this specialized knowledge and roll that out and scale that at a much bigger level than, um, you know, just one person speaking to another or trying to get it down in the playbook, be a bit more dynamic.

So I, I definitely see that coming to the forefront in the next, um, the next couple of years. Alright, well if there was one key takeaway you could leave our audience with today, what would that be? I mean, I think for me, I mean talking specifically to legal, um, but it, it does apply generally we've got to embrace the, the ai I think there's, there's, there's a natural hesitance.

I think the key takeaway for me is while there's risk and there's concerns, I think really understanding what like generative AI and these models can do is just of absolutely paramount importance. Everyone has to just understand the basics of how they, how it can benefit them, their team, and even the wider organization. Um, because it is life changing and it is, you know, for the job of a lawyer will not be the same again in a few years.

It just won't be, it's the, it is a big, big shift compared to what and what we've seen. So just spend time understanding how it works. The use cases, you know, I'm obviously biased 'cause we're a vendor that sells AI tools but like really kind of investigate AI tools and these, what we're seeing is huge compounding efficiencies that, that we are able to do now, um, using this technology that you don't want to be left behind.

And it's actually almost a, um, you know, a almost like a job satisfaction thing. It's a retention tool as well. So lean in, learn what AI can do, don't just focus on the risk and, and the governance side and um, and, and start experimenting with tools, you know, really, really understand what they can do for you.

Alright, well thank you so much for coming on the show and sharing your insights with us today. No, I appreciate you having me. All right.

And thank you to our audience. Stay tuned. There's more.