Episode 030 Workflow Options: Irys

WEBVTT
00:00:03.759 --> 00:00:06.799
Welcome to AI Tools for Practicing Lawyers.
00:00:07.040 --> 00:00:13.839
Practical, no-nonsense guidance to help attorneys put AI power to work in their practice right now.
00:00:14.960 --> 00:00:19.600
There's a strange thing happening in legal AI right now.
00:00:20.079 --> 00:00:24.079
On one hand, lawyers are building more than ever.
00:00:24.399 --> 00:00:26.239
We're building Claude skills.
00:00:26.320 --> 00:00:28.399
We're building custom GPTs.
00:00:28.559 --> 00:00:30.320
We're automating intake.
00:00:30.480 --> 00:00:42.000
We're connecting our documents and our email and trying to turn all of this incredible new technology into something that actually works in a law practice.
00:00:42.240 --> 00:00:51.039
But there's a pretty big difference between building a cool AI tool and building infrastructure.
00:00:51.280 --> 00:00:56.479
My guest today, Sabiz Siddiqui, knows that difference firsthand.
00:00:56.880 --> 00:01:08.319
He left Big Law to co-found Iris, a legal AI company that started with what looked like a fairly conventional off-the-shelf AI product.
00:01:08.640 --> 00:01:11.680
But then something interesting happened.
00:01:11.920 --> 00:01:14.480
Customers started asking for more.
00:01:14.879 --> 00:01:25.599
Iris started connecting systems, customizing workflows, and building deeper into the actual operations of law firms and legal departments.
00:01:25.840 --> 00:01:29.040
And that led us into a much bigger conversation.
00:01:29.439 --> 00:01:32.480
Should lawyers be building their own AI tools?
00:01:32.719 --> 00:01:34.239
Should they buy them?
00:01:34.560 --> 00:01:44.719
What happens when the big foundation models can suddenly do things that entire legal tech companies were built to do six months ago?
00:01:45.040 --> 00:01:49.359
Welcome to another episode of AI Tools for Practicing Lawyers.
00:01:49.439 --> 00:01:50.719
I'm Ron Drescher.
00:01:50.959 --> 00:01:52.000
I'm Heather Gardner.
00:01:52.560 --> 00:02:11.360
And our guest today is Sabi Siddiqui, a former big law litigator and outside general counsel who left behind the world of drafting NDAs and resolving disputes to become a legal tech builder and now CEO and co-founder of Iris.
00:02:11.840 --> 00:02:21.759
That's IRYS, a general-purpose legal AI tool that's offering a more accessible solution as opposed to the Harveys of the world.
00:02:22.080 --> 00:02:23.520
Sabi, welcome.
00:02:23.840 --> 00:02:25.280
Ron, Heather, thank you so much.
00:02:25.360 --> 00:02:26.560
And pleasure to be here.
00:02:26.800 --> 00:02:32.800
You know, before we went live, you said something that was very interesting.
00:02:32.960 --> 00:02:36.719
You said we're behind relative to AI.
00:02:36.879 --> 00:02:38.000
What did you mean by that?
00:02:38.319 --> 00:02:42.080
So, Ron, you know, as I was mentioning on a weekly basis, right?
00:02:42.240 --> 00:02:48.960
I go to conferences, talks here in New York City that have to do with legal tech and just AI generally.
00:02:49.199 --> 00:02:52.560
It's not necessarily that law lawyers are behind.
00:02:52.639 --> 00:02:55.199
It's that because, I mean, look at the nature of the business.
00:02:55.360 --> 00:02:57.360
We're very risk-averse people.
00:02:57.680 --> 00:02:58.800
We take our time with things.
00:02:58.879 --> 00:03:02.400
And obviously, tech isn't exactly top of mind for lawyers as well, either.
00:03:02.879 --> 00:03:16.080
So, what I mean by we're behind is in the sense of well, somewhat of a bit of adoption as well, but also understanding what AI means and what it can do for us as practitioners for the business.
00:03:16.240 --> 00:03:18.960
There's a lot of catching up that needs to happen.
00:03:19.199 --> 00:03:28.400
What was the data or the information that you were perceiving that caused you to conclude that we're behind?
00:03:29.120 --> 00:03:33.199
So there's more so the less so the data, more so the rhetoric.
00:03:33.439 --> 00:03:43.599
Just what's being talked about when I go to some of these conferences, it's still trying to understand whether it's the ROI, whether it's some of the more primitive data privacy concerns.
00:03:43.840 --> 00:03:46.639
Data privacy is a huge issue and aspect of it.
00:03:46.800 --> 00:03:55.840
But there's certain ways to think about this because there's been a lot of advancements, whether it's at the larger big AI model down to even our company, right?
00:03:56.000 --> 00:04:00.639
Whereas if you go to, let's say, an AI native conference, there's different conversations happening.
00:04:00.879 --> 00:04:09.360
So it's just a matter of at least people in my seat, people that are either founders or leading tech companies, to obviously educate the people that we're selling to.
00:04:09.680 --> 00:04:10.879
I don't fault lawyers for it.
00:04:11.039 --> 00:04:16.800
I think we're doing a phenomenal job and we're being rightfully cautious, but there is some catching up to do.
00:04:17.279 --> 00:04:26.720
You know, where because we do the podcast, Heather and I are seeing lawyers involved in tech in the AI world all the time.
00:04:26.879 --> 00:04:48.480
And it seems to us, or at least it seems to me, that there are so many lawyers out there that are like building skills in Claude, or, you know, trying to turn their law firms AI native, or going out like you did, to create AI tools in the legal field.
00:04:48.639 --> 00:04:50.639
Is that perspective skewed?
00:04:51.040 --> 00:04:53.839
I just last week was at a legal AI summit, as you know.
00:04:54.160 --> 00:05:08.160
So I am all I'm often at this at similar events that that you're discussing, Sabi, where we hear the rhetoric and the the doubts and the fears and just the there's something we've talked about many times on the podcast.
00:05:08.639 --> 00:05:16.319
Legal maybe a little behind, but we operate with abundance of caution in our industry for very valid reasons.
00:05:16.800 --> 00:05:19.199
But we are three years in now.
00:05:19.439 --> 00:05:21.920
So I understand where you're coming from.
00:05:22.160 --> 00:05:23.920
We're starting to really get behind.
00:05:24.079 --> 00:05:37.120
I think that it has been smart to have hesitation, to have caution, but we really do need to take advantage of and start learning and using these tools and helping more people, serving more consumers.
00:05:37.680 --> 00:05:39.199
I mean, Heather, that that's absolutely right.
00:05:39.279 --> 00:05:42.000
And Ron, to your point, it's the red car analogy.
00:05:42.480 --> 00:05:45.360
You talk or think about a red car, and then all you see is red cars.
00:05:45.600 --> 00:05:46.000
Yes.
00:05:46.160 --> 00:05:46.560
Yes.
00:05:46.879 --> 00:05:47.920
So kind of there.
00:05:48.079 --> 00:05:51.040
I mean, obviously, you have this phenomenal platform and podcasts.
00:05:51.199 --> 00:05:57.120
So I'm assuming that you're getting in some of the some of the tip of the top people that are doing exactly what you're talking about.
00:05:57.279 --> 00:06:05.759
But to to even double tap on one of the things that you'd mentioned, people, lawyers in particular, building on clawed, clawed skills and things of that nature.
00:06:06.000 --> 00:06:11.279
Well, we have to understand and ask ourselves as lawyers when we're doing something like that, what is going to be sustainable?
00:06:11.600 --> 00:06:14.079
At the end of the day, AI is moving so fast.
00:06:14.160 --> 00:06:16.800
And I'm saying that from the builder's side.
00:06:17.040 --> 00:06:23.199
The it the joke inside of our company is really every day in the real world is three days in our company.
00:06:23.439 --> 00:06:25.519
It's because something new is happening all the time.
00:06:25.759 --> 00:06:32.000
So, you know, we have to understand if you're creating a skill, okay, is Claude going to be the best place to actually deploy that?
00:06:32.160 --> 00:06:37.759
Can I have an environment in which, you know, from start to finish, I'm going to get the end result or work product?
00:06:37.920 --> 00:06:39.519
Or am I still going to have to be involved?
00:06:39.600 --> 00:06:41.759
Am I going to have to do certain checks and balances?
00:06:41.920 --> 00:06:45.920
And yes, as lawyers, we should always, always be checking the work.
00:06:46.079 --> 00:06:51.519
Um, but again, we have to think a little bit deeper about where, where and what we're building on.
00:06:51.759 --> 00:06:56.800
You know, I became a founder this summer, much to my chagrin.
00:06:57.120 --> 00:06:58.800
I never wanted to be a founder.
00:06:58.879 --> 00:07:00.720
I barely wanted to be a builder.
00:07:00.879 --> 00:07:18.720
And I discovered that there's a universe between building Claude skills, which are great to have in your practice, and actually trying to create a web app that's going to be available to lawyers just all over the world.
00:07:18.959 --> 00:07:25.439
How would you advise lawyers to bridge that gap if it's something they're really considering doing?
00:07:25.839 --> 00:07:30.720
So if it's something they're really considering doing, I'm not going to say, hey, come over to Iris.
00:07:31.519 --> 00:07:32.800
We are a solution.
00:07:32.879 --> 00:07:34.079
We're a phenomenal solution.
00:07:34.160 --> 00:07:36.319
And that's in the data that we can talk about, whenever.
00:07:36.560 --> 00:07:42.639
But at the end of the day, if you have that drive, that impetus, yes, try your hand at the skills.
00:07:42.800 --> 00:07:46.879
But just keep in mind, skills is it's a basic feature.
00:07:47.040 --> 00:07:52.560
What you're doing is you're providing Claude or an LLM a layer to operate on.
00:07:52.800 --> 00:08:02.319
Now, again, for example, we've just done last week a complete entire intake process for a team, a certain announced team.
00:08:02.480 --> 00:08:05.920
They had thought that, hey, we're using skills right now in Claude.
00:08:06.160 --> 00:08:07.439
In fact, they're on Claude.
00:08:07.680 --> 00:08:11.920
And they had thought that, hey, the skills can get us to where we need to get to.
00:08:12.319 --> 00:08:23.360
But for their intake with the systems, the integrations, while Claude had all the integrations, in fact, they still still were not able to piece together the right funnel or foundation, right?
00:08:23.600 --> 00:08:30.959
We put in 36 engineering hours, and now we're saving them close to, I think it's like 125 hours per month.
00:08:31.120 --> 00:08:32.879
And that's just a star, right?
00:08:33.039 --> 00:08:34.639
That's not what they were getting with Claude.
00:08:34.960 --> 00:08:38.159
That's because, again, we've built a whole ingestion pipeline.
00:08:38.320 --> 00:08:43.919
We have the memory, we have all of these support systems that are behind a real live application.
00:08:44.159 --> 00:08:48.960
So, okay, so maybe I had a misperception about your company.
00:08:49.120 --> 00:09:05.679
I got the feeling that Iris is an app that you buy off the shelf, you get your password, your login information, you pay your$299 a month, and you get to play around in this, you know, AI playground.
00:09:06.000 --> 00:09:11.679
But I'm hearing that it's that maybe your company is a little bit more hands-on with its customers.
00:09:11.919 --> 00:09:14.639
Or is it custom for each client?
00:09:15.120 --> 00:09:24.000
We have a bit of a dirty little secret where we started exactly with what you just mentioned, an off-the-shelf application.
00:09:24.240 --> 00:09:27.679
Now, the thing is, the way that we are built, think about it this way.
00:09:27.840 --> 00:09:37.600
Whereas if you have a, if you're a competitor, if you're building on top of a Chat GPT, you're building a certain scaffold, just like a New York City skyscraper.
00:09:37.679 --> 00:09:39.519
You're building on top of a foundation.
00:09:39.679 --> 00:09:44.240
And, you know, you there's some things that you just cannot move or you can't get off of.
00:09:44.480 --> 00:09:51.519
What we've done is we have think about it like Jenga blocks because it's our own infrastructure, it's our own harnesses.
00:09:51.679 --> 00:09:56.879
We're able to complete like actually be model agnostic, and we don't even start with the LLMs.
00:09:57.039 --> 00:10:01.759
So we can take things out, move things around pretty rapidly, extremely rapidly.
00:10:02.000 --> 00:10:09.759
So what we realized is that the legal platform, the legal tool that you're talking about, the 299, that has been an entryway.
00:10:10.000 --> 00:10:15.120
This specific in-house team that I'm talking about, I don't want to name them just Shep because we've done such a phenomenal job for them.
00:10:15.279 --> 00:10:17.759
I want there to be a larger splash around it.
00:10:18.000 --> 00:10:22.639
But um they started off with that 249 platform when it was 249.
00:10:22.720 --> 00:10:30.159
It was two that grew into their legal ops, then their compliance team, and now a bit of their accounting and finance teams are on it as well.
00:10:30.320 --> 00:10:33.200
We're talking about 77 people on the platform.
00:10:33.440 --> 00:10:43.039
So, and then as to Heather's point, what we did is we took the existing platform, white-labeled it slightly, but what we gave back to them is the existing product.
00:10:43.200 --> 00:10:45.759
So we productized this entire process.
00:10:45.919 --> 00:10:54.240
It's a bit of this palantier approach plus the new way of building, which if you have your own infrastructure, you can do whatever you really want.
00:10:54.320 --> 00:10:55.600
And that's what we are.
00:10:55.840 --> 00:11:01.039
How does that kind of conversation develop between a law firm and a vendor?
00:11:01.360 --> 00:11:06.639
They think they're buying something off the shelf for$249 a user a month and they're doing it.
00:11:06.799 --> 00:11:13.039
And how does it become apparent that, wait a second, there's a custom solution available for this company?
00:11:13.279 --> 00:11:15.200
It was a bit of discovery for both of us.
00:11:15.440 --> 00:11:18.000
And where did the discovery even start first?
00:11:18.399 --> 00:11:21.759
I practiced for close to eight years at the law firm.
00:11:21.840 --> 00:11:31.039
And, you know, even before, obviously, becoming a lawyer, before going to the firm, we're being sold to as law students, literally since day one.
00:11:31.360 --> 00:11:37.759
So we're used to a certain cadence, we're used to a certain sales cycle, a certain sales process, certain pitches, all that sort of other stuff.
00:11:38.080 --> 00:11:44.639
For whatever reason, even before the advent of AI, I was kind of part of some of our buying decisions as well.
00:11:44.799 --> 00:11:49.279
And obviously, I was getting emails, cold emails on a daily basis from certain vendors.
00:11:49.519 --> 00:11:51.360
I personally just hated all of it.
00:11:51.600 --> 00:11:53.360
Um I hated it for two reasons.
00:11:53.600 --> 00:11:57.039
I thought the tech generally was not that good.
00:11:57.200 --> 00:12:00.399
And then on top of that, there wasn't really an aspect of proper support.
00:12:00.480 --> 00:12:07.600
It felt like every time I spoke to one of the representatives from a company, my comments were falling on deaf ears.
00:12:07.759 --> 00:12:19.120
And in the back of my mind, I was thinking, like, damn, if they just did this one thing, if they just connected the dots here, this could be a whole item inside their business.
00:12:19.279 --> 00:12:20.879
This could be a business center.
00:12:21.120 --> 00:12:23.360
That's part of the reason also why I started the company.
00:12:23.600 --> 00:12:30.960
But what we realized is, hey, if you just be transparent, if you provide really good support, if you listen intently.
00:12:31.120 --> 00:12:36.799
And that's also one of the reasons why we didn't want to do this massive marketing campaign right in the beginning with our dollars.
00:12:37.039 --> 00:12:48.879
We wanted to grow in what we thought was a sustainable, very fast, but very sustainable way, so that the same exact gripes that people had with whether whether it's the legacy vendors or the newer vendors.
00:12:48.960 --> 00:12:53.759
And I'm, I know those newer vendors as well because again, I was part of the buying team at the firm before I left.
00:12:53.919 --> 00:12:55.919
But we didn't fall in, want to fall into that trap.
00:12:56.159 --> 00:13:02.559
We wanted to build something that actually mattered, that we could actually change practices.
00:13:02.720 --> 00:13:10.399
And that's happened on a daily basis with the limited, so to speak, clients we have granted, or close to 400 law firms and in-house teams now.
00:13:10.639 --> 00:13:15.200
We, if you, if a client emails me, like I'm gonna respond, right?
00:13:15.360 --> 00:13:18.879
Um, and if I can't respond, I'm forwarding that off to my support team.
00:13:19.039 --> 00:13:29.759
In our support channel on Slack, everyone has access to it, whether you're an engineer, whether you're on product, whether you're in sales, whether you're part of the support team, there's a phenomenal way in which we answer everything.
00:13:29.919 --> 00:13:30.799
We're getting you an answer.
00:13:31.039 --> 00:13:32.879
It's not gonna be bot that's responding either.
00:13:33.039 --> 00:13:39.200
So you're just talking about a customer base of about 400 different law firms.
00:13:39.440 --> 00:13:41.679
How did you penetrate that market?
00:13:42.320 --> 00:13:44.000
We exactly what I just mentioned.
00:13:44.080 --> 00:13:49.919
And and to build off of your earlier question, how did we, you know, grow from two to 77 or whatever?
00:13:50.240 --> 00:14:01.840
It was by offering that level of support, literally listening, like on the phone, someone calling them and understanding what they're what the pain points are, us understanding like, hey, we don't have to be just an application.
00:14:02.000 --> 00:14:03.360
We can move things around.
00:14:03.519 --> 00:14:09.679
And while we're moving things around, we can also then provide so much more to our existing clients as well.
00:14:09.840 --> 00:14:10.080
Right.
00:14:10.320 --> 00:14:16.080
So again, everything that that we're doing for someone might just go back into the product at the end of the day, unless it's truly, truly white-labeled.
00:14:16.240 --> 00:14:23.440
So that 400 law firm and in-house team penetration came quite literally from word of mouth.
00:14:23.600 --> 00:14:24.240
Nothing else.
00:14:24.399 --> 00:14:26.720
We did no marketing, no nothing.
00:14:26.960 --> 00:14:33.919
We, you know, I saw some of the Facebook groups that attorneys are in, some of the bar association um channels.
00:14:34.080 --> 00:14:35.440
That's where we were talked about.
00:14:35.600 --> 00:14:37.440
That's where we spread like wildfire.
00:14:37.679 --> 00:14:40.000
And it was just customers referring other customers.
00:14:40.159 --> 00:14:41.200
That's literally it.
00:14:41.440 --> 00:14:45.039
Only in the last few months we've spun up a GTM, a good market team.
00:14:45.840 --> 00:14:48.000
That sort of scale is just amazing.
00:14:48.240 --> 00:14:57.440
Are you are you in any particular practice area that stronger than others that enabled that kind of word of mouth?
00:14:57.679 --> 00:15:02.399
Ron, so I I think what you'll probably want from me is as much candor as possible.
00:15:02.559 --> 00:15:04.960
It seems like this this is the way it's yeah.
00:15:05.200 --> 00:15:07.120
So a little bit of an anecdote.
00:15:07.279 --> 00:15:21.440
I remember when Devanch, my co-founder and I, when we pitched Iris, and it was called a kidis back then, when we pitched Iris to the specific VC, there were like, guys, you gotta, you gotta specif specify as far as your practice area is concerned.
00:15:21.600 --> 00:15:25.039
Maybe do just plaintiffs, maybe just do defense work or whatever.
00:15:25.120 --> 00:15:28.320
But like you gotta be more specific than just this general thing you're talking about.
00:15:28.559 --> 00:15:33.759
Now, my co-founder, Devanch, is he's a phenomenal AI researcher and implementation engineer.
00:15:34.000 --> 00:15:36.240
He's literally always been ahead of the curve.
00:15:36.399 --> 00:15:46.639
And him and I together, we realized like, hell no, we're not just gonna go into concentrate on one specific practice area, grant it that would have done wonders, maybe even sold the company by now, who knows?
00:15:46.799 --> 00:15:47.279
Right.
00:15:47.600 --> 00:15:57.039
But the breadth of what we're doing and the breadth of what AI can do, if you take a look at Astra for law, it's not just focusing on, let's say, one specific or niche practice area.
00:15:57.840 --> 00:16:00.559
AI generally and models generally can do so much more.
00:16:00.720 --> 00:16:02.720
It depends on what you're feeding into the system.
00:16:02.960 --> 00:16:17.679
So, with us, if you are a law firm that's been in business for about five or 10 years, you have this plethora of data, you plug that into our system, give it a few hours for the maintenance and everything else to kick in, it'll come back to a very specific iris for you at the end of the day.
00:16:18.159 --> 00:16:27.360
And now with harnesses, the be the ability to call tools that previously weren't able to call, websites and things of that nature, concentrating on specific practice area, no.
00:16:27.519 --> 00:16:32.159
We've also been getting into some of the more IP stuff with T Tab litigation, stuff like that.
00:16:32.240 --> 00:16:34.080
But that's where our clients are taking us.
00:16:34.159 --> 00:16:37.519
And it's becoming very easy for us to do now, especially.
00:16:37.679 --> 00:16:48.159
So do you think the mega companies like Harvey and Lagora had to answer the same questions when they went to get their VC funding?
00:16:48.480 --> 00:16:54.799
Because they they're a general, they're both general purpose AI legal tools.
00:16:55.120 --> 00:17:03.360
So we've we've we've spoken to some of their teams, not only from a buyer's side, but I mean, literally spoken to either engineering, we've spoken to their investors.
00:17:03.840 --> 00:17:07.119
We understand what and why they did and how they did it.
00:17:07.440 --> 00:17:11.680
There is a huge or was a huge, huge first mover advantage.
00:17:11.839 --> 00:17:16.559
They did exactly the right thing that I would have done if I was first to move like they were.
00:17:16.720 --> 00:17:19.440
We are wave two, and we'll talk about what wave two is.
00:17:19.680 --> 00:17:26.799
They, I think, had a leg up in the sense of their Harvey specifically, OpenAI was one of their first investors.
00:17:27.440 --> 00:17:31.519
GPT generally is not specific to any one practice area, so to speak.
00:17:31.759 --> 00:17:36.559
So they were likely able to show that, hey, this doesn't apply to just one thing or place.
00:17:36.720 --> 00:17:38.079
We can do so much more.
00:17:38.240 --> 00:17:42.559
Now, are they answering the same questions that we are answering right now for our clients?
00:17:42.720 --> 00:17:43.440
Yes and no.
00:17:43.680 --> 00:17:48.079
I'll be honest, and both of those teams know this as well because we won their clients over.
00:17:48.240 --> 00:17:53.359
Um, it's that the way that we're able to work is is deeper than them because of how we're built.
00:17:53.519 --> 00:17:54.960
They have done a phenomenal job.
00:17:55.039 --> 00:17:56.960
They're not bad companies whatsoever.
00:17:57.119 --> 00:18:00.960
We are just very different, and we're also going in a very different direction right now.
00:18:01.119 --> 00:18:09.440
So listen, the the tech at the end of the day, the underlying infrastructure, granted, if you open the 299 application, right, it might look very similar.
00:18:09.519 --> 00:18:12.319
And a lot of attorneys say, like, holy, oh wow, this is better.
00:18:12.480 --> 00:18:12.960
Yes, it is.
00:18:13.039 --> 00:18:15.279
It's built by lawyers, and we're actually listening to our customers.
00:18:15.440 --> 00:18:15.759
Great.
00:18:15.920 --> 00:18:24.000
But at the end of the day, the ingestion, the memory, the permissions, all of that is vastly different because we're not just, we're not starting with the models first.
00:18:24.160 --> 00:18:30.079
We're starting with our own pipeline and then bring in the models to do the job as far as the natural language processing is concerned.
00:18:30.319 --> 00:18:32.079
That is the last step in what we do.
00:18:32.240 --> 00:18:36.400
So the answers tend to be vastly different or better.
00:18:36.559 --> 00:18:43.279
The amount of information you're able to put in and have really good quality output is because we have our own systems.
00:18:43.359 --> 00:18:44.720
That's the underlying tech.
00:18:44.960 --> 00:18:51.599
What we're doing right now, as I mentioned earlier, it's not that we're pivoting, it's that we're fitting into our shoes now.
00:18:51.680 --> 00:18:54.559
And our shoes are not just an application layer.
00:18:54.720 --> 00:19:02.720
Our shoes truly are you start with the application, but then you do all these deployments at large scales and at cutting edge cost.
00:19:03.039 --> 00:19:11.680
What I mean by that, for example, is this specific company that I was talking about, they had earmarked certain, they they thought it was going to cost them a certain amount.
00:19:11.839 --> 00:19:19.599
They thought it was gonna take a certain amount of time because they had gone and done their diligence with some larger competitors for this custom intake that I'm talking about.
00:19:19.759 --> 00:19:25.519
We came in at 75% of the cost and 90% 90% of the time, 36 engineering hours versus whatever.
00:19:25.839 --> 00:19:29.920
I've discussed this with other lawyers, both on the show and off the show.
00:19:30.160 --> 00:19:43.359
And it seems to me that the legal field is racing at a breakneck pace to adopt AI as kind of part of its overall presentation.
00:19:43.440 --> 00:19:47.920
I mean, look, I see you're you've got a little skepticism, and I I respect that.
00:19:48.240 --> 00:20:01.200
But at the same time, every CLE in the United States in the last six months and the next 18 months is going to be about AI.
00:20:01.599 --> 00:20:10.640
Every single, you know, and in fact, in you know, I'm looking at other CLEs like uh, you know, we're we're Heather and I are both really bankruptcy guys.
00:20:10.720 --> 00:20:18.000
And I saw a CLE for subchapter V bankruptcies, and I'm saying, who's going to that?
00:20:18.240 --> 00:20:21.119
Everybody's going to the AI CLEs.
00:20:21.359 --> 00:20:34.960
But so that's what I'm perceiving, which I think is counterintuitive because lawyers are so risk averse, and AI carries with it so many, you know, fundamental risks.
00:20:35.119 --> 00:20:40.480
How do we reconcile those two kind of seemingly disparate trends?
00:20:40.720 --> 00:20:45.440
And I want to just clarify because I'm doing a lot of those, you know, CLEs.
00:20:45.519 --> 00:20:49.680
I'm on these panels, and I just want to clarify they're racing to the CLEs.
00:20:49.920 --> 00:20:54.720
They're not actually all implementing it in their offices, and they are still scared to.
00:20:54.880 --> 00:20:58.799
And we've had so many attorneys on that have talked about a lot of the reasons.
00:20:58.880 --> 00:21:10.400
And as we've talked to several bankruptcy attorneys who don't want to buy one off-the-shelf thing, what he's talking about is offering something that is custom, which seems to be what attorneys want.
00:21:10.960 --> 00:21:11.200
Yeah.
00:21:11.359 --> 00:21:14.240
So, I mean, Heather and Ron, excellent, excellent question.
00:21:14.480 --> 00:21:15.119
Why?
00:21:15.519 --> 00:21:21.599
Because it depends on kind of who you're talking to as well, whether it's an N-house team, whether it's lawyers at a law firm.
00:21:21.839 --> 00:21:29.599
What we have realized at the end of the day, it comes down to the business of law, which is what is scaring people the most.
00:21:29.759 --> 00:21:37.920
When we speak to partners, whether they're at a AM Law 100 or mid-sized or even smaller firms, boutiques, it's it's about the ROI.
00:21:38.079 --> 00:21:40.319
If we talk to an N-house team, it's about the ROI.
00:21:40.480 --> 00:21:44.559
The ROI on the in-house side is much easier to realize.
00:21:44.720 --> 00:21:45.119
Why?
00:21:45.279 --> 00:21:54.960
Because as I mentioned before, sticking with that one same example, if this team is saving 100 plus hours, they're also saving a quarter of that that's that was going to their outside council.
00:21:55.119 --> 00:21:55.440
Okay.
00:21:55.599 --> 00:21:57.680
So the ROI for them is huge.
00:21:57.839 --> 00:22:01.839
And the best part is what they're doing is They're not hiring less than this NOS team.
00:22:01.920 --> 00:22:04.880
They're actually going to hire a couple more attorneys to join their team.
00:22:05.039 --> 00:22:05.440
Why?
00:22:05.680 --> 00:22:11.440
Because now those attorneys, much like our engineers and our team, are 100x, what they could do before.
00:22:12.000 --> 00:22:28.400
So while they still may be viewed as cost centers, if they are saving the larger organization money on the tail line, whether it's outside council fees, whether it's being able to do processes faster or get to more deals, that's that's a phenomenal place to be.
00:22:28.559 --> 00:22:30.079
So it comes down to the ROI.
00:22:30.240 --> 00:22:35.440
There are absolutely risks with privacy and with things like, you know, hallucinations and stuff like that.
00:22:35.519 --> 00:22:36.960
We can talk about that as well.
00:22:37.119 --> 00:22:44.960
But the reason that we have seen that people are scared really is the ROI and then the hesitation to adopt something new or do something new.
00:22:45.200 --> 00:22:47.039
That's really always.
00:22:47.200 --> 00:22:47.759
Yeah.
00:22:48.079 --> 00:22:54.319
And and not to drone on, but you know, I literally remember before I left the firm, while I did dispute res.
00:22:54.640 --> 00:22:56.799
Yes, I was outside Council World Bank as well.
00:22:56.960 --> 00:23:00.160
For this one larger client that had a subsidiary that was going into bankruptcy.
00:23:00.480 --> 00:23:00.960
It was weird.
00:23:01.119 --> 00:23:04.240
We were thinking about liquidation forms as well, chapter seven, stuff like that.
00:23:04.319 --> 00:23:05.759
So I was doing some of those forms.
00:23:05.920 --> 00:23:10.559
I had used a particular platform, an AI platform to actually do those forms.
00:23:10.640 --> 00:23:13.599
And I was like, shit, that would have taken me at least 45 minutes.
00:23:13.680 --> 00:23:14.400
But guess what?
00:23:14.480 --> 00:23:15.440
It took me 10.
00:23:15.680 --> 00:23:18.160
I saw the writing on the wall up until the day that I left.
00:23:18.319 --> 00:23:22.079
So it's a matter of how we want to integrate AI into our practice.
00:23:22.240 --> 00:23:25.279
We still have not figured that out on a law firm side entirely.
00:23:25.440 --> 00:23:33.599
What we have seen anecdotally and tangibly from our own clients is if you're doing things like opinion letters, things that are more routine or based on precedent.
00:23:33.680 --> 00:23:40.960
And let's say you get an email or you have this an agreement that you got to base this new opinion letter on, that's very easy to do on a fee basis.
00:23:41.119 --> 00:23:47.119
So a client of ours that did that on an hourly basis, now doing it for fee, they charge a flat rate of 3,500 bucks.
00:23:47.279 --> 00:23:56.960
It takes them no more than one hour to do these opinion letters, but they're the clients are paying for this firm's years, decades of judgment that's going into this opinion letter.
00:23:57.200 --> 00:24:00.880
So we're not necessarily seeing them lose money, but they're not losing money whatsoever on this.
00:24:01.039 --> 00:24:02.720
It's taking them one hour max.
00:24:03.039 --> 00:24:08.640
But it's having the wherewithal to actually take a step back and understand what is it that I'm providing?
00:24:08.799 --> 00:24:10.319
What are the services that I'm providing?
00:24:10.480 --> 00:24:12.880
What can I productize as a law firm?
00:24:13.119 --> 00:24:14.559
Because I'm providing an outcome.
00:24:14.720 --> 00:24:15.680
Can that be a product?
00:24:15.839 --> 00:24:16.559
Absolutely.
00:24:16.799 --> 00:24:18.559
It's a it's a mindset shift.
00:24:18.720 --> 00:24:21.920
And I had a ton of my own business before I left the firm as well.
00:24:22.079 --> 00:24:29.039
One of the bigger reasons why I left was not just because I wanted to do product, but I saw the inefficiencies firsthand.
00:24:29.200 --> 00:24:32.720
Was I always worth, you know, whatever, 1,100 bucks an hour or whatever?
00:24:32.960 --> 00:24:33.279
Sure.
00:24:33.440 --> 00:24:33.680
Right?
00:24:33.839 --> 00:24:36.160
Is that always I don't think so.
00:24:36.319 --> 00:24:36.640
Right.
00:24:36.799 --> 00:24:41.599
But now By the way, I I think there are babysitters that are worth$1,100 an hour.
00:24:41.839 --> 00:24:44.559
I think teachers should get a million dollars an hour.
00:24:44.640 --> 00:24:45.200
I'm just saying.
00:24:45.680 --> 00:24:47.279
I completely agree with it.
00:24:47.440 --> 00:24:48.319
I completely agree.
00:24:48.559 --> 00:24:49.039
Right.
00:24:49.279 --> 00:25:01.759
But, you know, if if AI is delivered in such a way that maybe it's our company, maybe it's not, but we really do want to get to the core of the business and practice of law.
00:25:01.839 --> 00:25:02.240
Why?
00:25:02.400 --> 00:25:03.839
Because we think that it can change.
00:25:04.079 --> 00:25:12.079
We know it's not going to be an overnight change, but for the betterment of society or whatever you want to call it, like that is something that's actually really top of mind for us.
00:25:12.319 --> 00:25:15.200
The efficiency can be value-driven.
00:25:15.359 --> 00:25:17.680
That can provide a change in pricing.
00:25:17.920 --> 00:25:24.880
We have seen our attorney saying that, hey, we're not working less, but we're doing so much more better work for our clients, which everyone's really happy about.
00:25:25.119 --> 00:25:25.759
So yeah.
00:25:26.079 --> 00:25:52.559
You know, I was talking to an attorney uh yesterday, the day before, and the attorney said, this attorney from a big law firm, and said, Well, our in-house counsel clients, they they know we use AI, they want us to use AI, they insist upon us using AI, and then they insist upon us documenting how we use AI and how that's saving them money.
00:25:52.799 --> 00:25:53.440
Yeah.
00:25:53.759 --> 00:25:59.359
How would you advise that lawyer to implement a response to that demand?
00:25:59.680 --> 00:26:02.480
And does your company provide those metrics for your clients?
00:26:02.720 --> 00:26:02.880
Yeah.
00:26:02.960 --> 00:26:04.480
So even better.
00:26:04.640 --> 00:26:04.799
Yeah.
00:26:05.039 --> 00:26:12.160
But that's, and again, I do not want to be just plugging, but there's so much anecdotal evidence for what I'm going to say right now.
00:26:12.400 --> 00:26:16.400
Another N-house team that's on our platform, they're a fintech company.
00:26:16.960 --> 00:26:24.960
They basically asked us, it's September right now, earlier in the year, hey, you know, we want to be able to connect with our law firms.
00:26:25.359 --> 00:26:27.200
We want them to start adopting AI.
00:26:27.359 --> 00:26:29.920
We want to understand how they're using it because we're using it.
00:26:30.079 --> 00:26:32.240
We're still getting the same level of bills, though.
00:26:32.400 --> 00:26:34.240
They're saying they're using it, but we don't know.
00:26:34.480 --> 00:26:36.240
Can you help us in some ways?
00:26:36.480 --> 00:26:39.279
Funny enough, I talked to my sales team about this.
00:26:39.519 --> 00:26:42.640
And we we might have even missed this if we didn't go so deep.
00:26:42.880 --> 00:26:48.000
But one of their outside counsels is or was and still is on the platform.
00:26:48.559 --> 00:26:54.720
So it was, we had, we wanted to broach the conversation, like, hey, can we bring the parties together here?
00:26:55.119 --> 00:27:07.200
And what we did for them is we actually created a bridge where this fintech company that does a ton of regulatory reports specifically with this one outside firm, it's a very, it's a relatively routine process.
00:27:07.519 --> 00:27:21.200
And giving them that bridge, basically a client portal on steroids, allow them to understand how this specific firm is using AI because, again, they're literally on the same platform with them.
00:27:21.519 --> 00:27:29.440
So that is, I think, the most ideal way to do this, where the in-house teams and lawyers are in the same workspace, actually.
00:27:29.680 --> 00:27:38.960
Otherwise, what we've also done, and Heather, to your point, we send month end, if you are an enterprise client, your basically ROI report.
00:27:39.119 --> 00:27:40.799
We literally call it an ROI report.
00:27:40.960 --> 00:27:52.559
And what we've now done for three larger clients, and because they have like droves of their attorneys on it and practice groups, because practice groups, we're able to break it down by practice groups.
00:27:52.720 --> 00:28:00.319
Now, what we did as a pilot with one specific firm, the way that we've set up the platform, the online platform, is you're working in matters.
00:28:00.480 --> 00:28:04.400
So whether you're at an in-house or or at a law firm, you can work in a matter.
00:28:04.559 --> 00:28:07.920
Because they're working in matters, they've organized all their matters in our client folders.
00:28:08.000 --> 00:28:11.279
We were able to give ROI reports per client folder, basically.
00:28:11.440 --> 00:28:11.759
Right.
00:28:11.920 --> 00:28:14.400
Um, that has been a hit for the specific firm.
00:28:14.720 --> 00:28:17.519
Query as to whether every single firm is going to want that.
00:28:17.680 --> 00:28:22.400
The thing also is we're kind of a little, our hands are kind of tied right now.
00:28:22.559 --> 00:28:25.119
We've grown faster than we anticipated.
00:28:25.279 --> 00:28:33.039
So part of the reason of what we're we're going to be doing with our with our upcoming raise and stuff is literally just getting more support, people, bodies.
00:28:33.200 --> 00:28:36.160
Not everything can be AI-ified whatsoever.
00:28:36.319 --> 00:28:38.000
You still need that human touch.
00:28:38.079 --> 00:28:39.119
That's not going to go away.
00:28:39.279 --> 00:28:41.759
So what can't what can't be AI-ified?
00:28:42.000 --> 00:28:42.880
Judgment.
00:28:43.279 --> 00:28:43.839
Judgment.
00:28:44.000 --> 00:28:46.480
Human judgment cannot be AI-ified.
00:28:46.880 --> 00:28:48.640
AI can give you reasoning.
00:28:48.720 --> 00:28:52.720
It can give you a output on the basis of what's been inputted.
00:28:52.960 --> 00:28:58.160
However, human reasoning and judgment is just vastly different, and that will not be AI-fied.
00:28:58.400 --> 00:29:02.480
That is what a client is paying SCADN to do, right?
00:29:02.640 --> 00:29:05.599
And that is what a client is paying any other big law firm to do.
00:29:05.759 --> 00:29:10.880
AI can assist to getting to that judgment quicker, better, being more informed.
00:29:11.039 --> 00:29:15.599
But that judgment is what AI is not going to do away with at the end of the day.
00:29:16.079 --> 00:29:20.319
Let's talk about dollars because your price point is so impressive.
00:29:21.200 --> 00:29:27.200
We look when we talk about the gap between big law and law firms having access to tech like this.
00:29:27.759 --> 00:29:36.240
Starting at$299 a month, your team starting at$329 a month, Harvey, I believe, is what,$1,200 to$1,500 a month per user, and you have to have 100 seats or something of that nature.
00:29:36.799 --> 00:29:38.799
No, no, they don't go public with that.
00:29:39.039 --> 00:29:39.279
Right.
00:29:39.359 --> 00:29:40.319
But I you can find it.
00:29:40.480 --> 00:29:40.960
You can find it.
00:29:41.039 --> 00:29:41.519
You can find it.
00:29:41.599 --> 00:29:42.319
I found that information.
00:29:42.720 --> 00:29:43.759
Yeah, I've I found that information.
00:29:44.960 --> 00:29:45.519
What was that?
00:29:45.680 --> 00:29:47.599
So to add a little bit of color.
00:29:47.759 --> 00:29:50.640
The I don't know how much trouble I'm going to get in for this, but whatever.
00:29:50.720 --> 00:29:52.559
We'll take it on the chin if we didn't.
00:29:52.799 --> 00:29:56.960
Obviously, through clients and other people, we we know the pricing of Harvey and what it used to be.
00:29:57.119 --> 00:30:01.680
At least it was$1,200 to$2,600, depending on if you had the Lexus integration.
00:30:02.160 --> 00:30:05.920
At the same time, both Harvey and Lagora have been going down market.
00:30:06.400 --> 00:30:16.079
However, you know, the the cautionary tale for all of that is if you take a look at the customer agreements, what they're doing is they're doing two things.
00:30:16.319 --> 00:30:20.720
They're moving certain customers and segments onto lower tier models.
00:30:20.880 --> 00:30:22.480
So you're not getting the best intelligence.
00:30:22.640 --> 00:30:30.160
And if you're wrapping around another model and that's where you're providing us intelligence, that just means you're getting lower quality responses at the end of the day.
00:30:30.400 --> 00:30:36.640
That's not to, you know, we're not trying to slap them on the we're not we're not trying to bash anyone, but that's just the reality of the situation.
00:30:36.799 --> 00:30:39.279
That's why the smaller firms are kind of mad.
00:30:39.519 --> 00:30:44.319
On top of that, what they're doing is uh they're use, they're moving to usage-based pricing.
00:30:44.480 --> 00:30:50.720
A lot of companies are saying, oh, usage-based pricing is good because maybe not all of your, you're not all of your users are going to use it.
00:30:50.799 --> 00:30:52.480
So just pay us this nominal fee.
00:30:52.640 --> 00:30:56.400
One of our clients, I mean, you take you can take a look at the Claude anecdote that we had.
00:30:56.559 --> 00:31:01.839
One client that, you know, seven of their attorneys were using Iris, two of their attorneys were using Claude.
00:31:02.079 --> 00:31:04.559
They ran up a 4K bill in two weeks using Claude.
00:31:04.880 --> 00:31:05.119
Totally.
00:31:05.359 --> 00:31:08.000
Their bill on Iris was their bill.
00:31:08.079 --> 00:31:08.319
That's it.
00:31:09.119 --> 00:31:10.640
So it wasn't, that wasn't the question, though.
00:31:10.720 --> 00:31:14.079
This my question wasn't how much does it cost versus Harvey?
00:31:14.240 --> 00:31:15.440
Because the the gap is there.
00:31:15.599 --> 00:31:16.960
Of course, we know that it's huge.
00:31:17.119 --> 00:31:28.079
My question was when you're doing a custom build, which is what we what I think is amazing because every attorney we have on the show talks about, they don't want the off-the-shelf AI tool.
00:31:28.240 --> 00:31:29.119
They want their own.
00:31:29.279 --> 00:31:31.359
That's why they're all building their own quad skills.
00:31:31.440 --> 00:31:35.759
That's why they're building their own verticals, because they want what they want.
00:31:36.000 --> 00:31:42.319
So if you provide that, that's gonna be a that's gonna be a, you know, which we've had another vendor on that provides a similar thing.
00:31:42.480 --> 00:31:47.359
That is going to be what I feel like is the most valuable AI tools for attorneys.
00:31:47.599 --> 00:31:54.400
But if you're building out that custom build, are you getting a lot closer to that Harby or that uh that monthly?
00:31:54.559 --> 00:32:00.000
Is because I'm guessing that's not gonna be in your$329 a month per se range.
00:32:00.240 --> 00:32:12.799
Yeah, it it really depends because even what Ron had mentioned earlier, the skills can certainly, and we have a phenomenal workflow and skills library, that can get you really, really far.
00:32:13.680 --> 00:32:24.880
But if you have disparate systems, if you are, for example, there's one larger firm we're working with, they have a I don't know if I can name the DMS, they have a DMS provider.
00:32:25.119 --> 00:32:28.160
However, they have their documents in other places as well.
00:32:28.480 --> 00:32:29.599
That's not easy to do.
00:32:29.759 --> 00:32:30.640
No one can do that.
00:32:31.039 --> 00:32:34.400
We can do that because it's an easy ingestion question for us.
00:32:34.559 --> 00:32:44.960
So the way that we price it for this specific firm is you know, the sales team plus engineering plus product worked out a number and that came out to$399 per month for this specific client.
00:32:45.200 --> 00:32:52.319
We didn't even basically we said, like, hey, we know quite honestly, this is going to take us X amount of hours.
00:32:52.480 --> 00:32:54.640
It's not exactly the end of the world for us.
00:32:54.880 --> 00:32:59.200
However, there's gonna be maintenance costs on the back end because we're tying together these systems.
00:32:59.440 --> 00:33:02.000
This is the price that we think is gonna be best suited for this job.
00:33:02.160 --> 00:33:05.039
Let's do an annual contract and then go forward from there.
00:33:05.359 --> 00:33:06.720
That was a custom build?
00:33:07.039 --> 00:33:07.599
Custom build.
00:33:07.839 --> 00:33:08.720
That was a custom build.
00:33:08.799 --> 00:33:08.960
Okay.
00:33:09.119 --> 00:33:09.359
Yeah.
00:33:09.599 --> 00:33:23.759
You know, I I'm not so sure I agree, Heather, that I mean, yes, the attorneys that we get on the show all want to build their own stuff because they've already kind of established their place as a, you know, kind of an AI maven.
00:33:24.240 --> 00:33:37.359
But is that really true amongst so many of the rank and file lawyers that you sh that show up in court, you know, every morning for their 30-second hearings and and then go back to their office?
00:33:37.519 --> 00:33:42.880
But I I kind of get the sense that lawyers would be more than happy to get a solution that's off the shelf.
00:33:43.119 --> 00:33:43.599
Absolutely.
00:33:43.759 --> 00:33:44.079
Right?
00:33:44.160 --> 00:33:49.599
Specifically, if we were talking about mid-size and smaller boutiques, it's a time thing.
00:33:49.759 --> 00:33:50.960
And they're busy as hell.
00:33:51.039 --> 00:33:53.119
They're providing value on a daily basis for their clients.
00:33:53.200 --> 00:33:56.000
So they don't want to get involved with whether it's custom or even talking.
00:33:56.079 --> 00:33:58.720
There's skepticism involved, a lot of skepticism.
00:33:58.799 --> 00:34:01.279
Um, and then the pricing is always a concern as well.
00:34:01.440 --> 00:34:04.640
So the$299 gets them across the board or across the shelf.
00:34:04.799 --> 00:34:12.480
And then on top of that, with the support that we offer literally everyone, knock on wood, and hopefully we were able to continuously offer that to everyone, they're really able to get the job done.
00:34:12.639 --> 00:34:21.039
So yeah, the hundreds of attorneys I've worked with in the past five years, and you know, many, many before the past five years, they all want what they want.
00:34:21.199 --> 00:34:27.760
And just like we had that quote, uh, Teris Salinas said that I've never met an attorney that didn't want things done their way.
00:34:28.000 --> 00:34:37.119
So I do think that if they had the option to have someone build exactly what they want versus something off the shelf, and you're looking at a similar price point, it seems like a no-brainer.
00:34:37.679 --> 00:34:38.159
It does.
00:34:38.480 --> 00:34:48.480
And but so one of the reasons why I started Iris was it felt continuously like, okay, you have these bigger companies out there.
00:34:48.639 --> 00:34:52.400
Somehow, some way it just feels like was rewards actually behind this.
00:34:53.360 --> 00:34:57.840
The magic was lost as far as the tech is concerned because it was no different than ChatGPT, quite frankly.
00:34:58.559 --> 00:35:07.039
So, what we have been continuously trying to do, specifically with the off-the-shelf option, is look into the future, connect the dots.
00:35:07.199 --> 00:35:08.480
What do people want today?
00:35:08.559 --> 00:35:10.000
But what do they really want tomorrow?
00:35:10.320 --> 00:35:12.000
How can we make the practice better tomorrow?
00:35:12.079 --> 00:35:15.199
Because we're going to be able to do the base level functionality hands down.
00:35:15.519 --> 00:35:25.760
But for example, like even with button placement, even with things like the amplified prompt, even with things like how you can customize the jurisdiction and stuff like that.
00:35:26.079 --> 00:35:29.280
We're paying attention to all of that really intently.
00:35:29.679 --> 00:35:39.679
And I I I can't, I really can't say that for everyone else, but it it comes down to also understanding and listening on a deeper level that deeply deeper level for some reason.
00:35:39.760 --> 00:35:40.320
So yeah.
00:35:40.639 --> 00:35:50.800
You're you're building your you're you're building your product and your company the exact way AI was built by learning pattern recognition and dumping information into it.
00:35:50.960 --> 00:35:52.480
That is genius, actually.
00:35:53.280 --> 00:35:54.559
And it's what I would have wanted.
00:35:54.719 --> 00:35:54.960
Right.
00:35:55.039 --> 00:35:56.880
It's really what I would have wanted.
00:35:57.039 --> 00:36:12.800
So one of the things that I've really appreciated about this movement in the AI world is that it has caused lawyers and law firms to focus on their systems and their processes for getting work done and for for managing their workflow.
00:36:13.039 --> 00:36:15.920
And I I want to get your take on this.
00:36:16.079 --> 00:36:29.760
All other things being equal, do you think a firm is going to be better off actually implementing some AI solution or actually having its lawyers and staff learn the tools that are already in-house?
00:36:30.000 --> 00:36:34.719
The phone system, Microsoft Word, all the features of Slack.
00:36:34.880 --> 00:36:37.840
Where do you think the best ROI is there?
00:36:38.239 --> 00:36:41.119
So, Ron, that is it's such a phenomenal question.
00:36:41.280 --> 00:36:41.599
Why?
00:36:41.840 --> 00:36:43.360
Because I was that guy.
00:36:43.599 --> 00:36:44.559
He was the one.
00:36:44.639 --> 00:36:49.360
And I was part of the new innovation team that we had launched at the firm before I left as well.
00:36:49.599 --> 00:36:54.880
But for years prior to that, I wanted to take as much advantage of the tech that was available to us at that time.
00:36:55.280 --> 00:37:02.079
Frankly, the the problem is that even with Slack, whether it's Skype, Microsoft, Team, Zoom, everyone's going on the AI route as well.
00:37:02.159 --> 00:37:02.400
Sure.
00:37:02.480 --> 00:37:09.679
So there's some learning with AI, but it's still the efficiency aspect, those are tools.
00:37:09.920 --> 00:37:10.239
Okay.
00:37:10.639 --> 00:37:12.480
Tools can only do so much.
00:37:12.719 --> 00:37:16.880
A platform that unifies the way that you work, that's a different story.
00:37:17.119 --> 00:37:26.000
So you can become better and you can become slightly more efficient and deliver some more value to your clients by learning the tools, such as Skype or whatever it is.
00:37:26.079 --> 00:37:31.199
I'm dating myself, maybe a little bit, but the platforms, I think, really is the way to go.
00:37:31.440 --> 00:37:31.679
All right.
00:37:31.760 --> 00:37:34.960
Well, Sabin, this is just a wonderful conversation.
00:37:35.199 --> 00:37:45.119
I have, I have a list of about 50 questions I haven't gotten to, but uh I think it's time for us to move over to our practice signal segments.
00:37:45.280 --> 00:37:54.559
And this is a segment where we look at uh how lawyers in the wild are struggling and maybe potential solutions we have for them.
00:37:54.639 --> 00:37:58.400
And maybe that there's an AI solution, maybe there isn't an AI solution.
00:37:58.559 --> 00:38:05.840
They always come from Reddit because, you know, you can't have more fun than you can have, you know, scrolling through Reddit.
00:38:06.000 --> 00:38:06.719
For sure.
00:38:06.960 --> 00:38:08.960
So so here, this one comes.
00:38:09.039 --> 00:38:16.079
Uh, we're trying to find a better process for opening new cases with our PI.
00:38:16.400 --> 00:38:22.480
Currently, we have a few people in the office that help each other out with the different stages of a new case.
00:38:22.639 --> 00:38:28.400
But as we get busier, this has created gaps in who owns the next step of the file.
00:38:28.639 --> 00:38:31.199
See, Sub Bee, this is going to be a softball for you.
00:38:31.360 --> 00:38:45.360
Would love to hear how others have found a good way to streamline process and ensure that a new case is opened efficiently, quickly, and without unnecessary bottlenecks.
00:38:45.760 --> 00:39:00.159
So if you walked into that firm tomorrow, Sub Bee, how would you diagnose the problem and what would Iris or any well-designed system actually do to make ownership, deadlines, and the next required action unmistakable?
00:39:00.400 --> 00:39:03.599
Let me do this because I I won't be so self-serving.
00:39:03.840 --> 00:39:07.519
If I were to walk in tomorrow and I had no affiliation with Iris, right?
00:39:07.599 --> 00:39:18.480
And I just knew AI really, really well, what I would do is I would set up Claude or ChatGPT projects and I would have an integration to ideally Gmail.
00:39:18.559 --> 00:39:24.400
Um, and this firm would ideally, because I'm using ChatGPT or just Claude, there would be a G drive team.
00:39:24.559 --> 00:39:26.079
That would make it a lot easier.
00:39:26.320 --> 00:39:29.599
That way I can set up a process for that intake.
00:39:29.760 --> 00:39:32.800
It's going to require a little bit more of a manual effort with GPT.
00:39:32.880 --> 00:39:38.880
It can be a little bit easier with skills on Claude and routines that may start running up a bill.
00:39:39.039 --> 00:39:46.239
I mean, not may, it will definitely start running up a bill, specifically if you are prosecuting all of those PI cases within Claude.
00:39:46.480 --> 00:39:59.920
It doesn't make sense to break up the system where you're doing work off of Claude or you're at least not feeding in your documents and drafts to Claude, because then you're not going to be able to go from beginning to end thoroughly, right?
00:40:00.000 --> 00:40:01.840
So you're going to want to be doing all that work there.
00:40:02.000 --> 00:40:06.400
Now, a system like Iris, which charges by C, obviously you're able to do all of that.
00:40:06.559 --> 00:40:18.320
And then the best part is not only are we releasing an intake, a whole new intake system, literally within the next 14 days, but we are already set up to basically do most of what I just talked about, right?
00:40:18.480 --> 00:40:21.039
And you're not going to be charged on tokens.
00:40:21.199 --> 00:40:23.599
So you can set up a matter automatically.
00:40:23.760 --> 00:40:26.880
You can assign your team members to that matter or that client.
00:40:27.119 --> 00:40:35.280
So there is going to be literally something for every single person at this firm, whether you are a paralegal, junior, senior associate, or partner.
00:40:35.599 --> 00:40:37.440
Partner has visibility into the matter.
00:40:37.679 --> 00:40:41.280
A senior and junior associate can do the work along with the paralegal as well.
00:40:41.440 --> 00:40:44.480
Everyone's in there setting up tasks, deadlines, things of that nature.
00:40:44.639 --> 00:40:48.559
And we made it extremely simple and easy for the whole team to be on the platform.
00:40:48.719 --> 00:40:50.400
That's what I would do if I knew about Iris.
00:40:50.559 --> 00:40:53.840
If I didn't know about Iris, I would set it up on Claude.
00:40:54.159 --> 00:41:00.639
You know, uh Claude has become the favorite of the legal builder community.
00:41:00.880 --> 00:41:08.559
OpenAI just came out with Astra for legal, uh, about four months after Claude came out for Claude for Legal.
00:41:08.800 --> 00:41:12.239
Is ChatGPT have they caught up?
00:41:12.480 --> 00:41:13.920
Will they ever catch up?
00:41:14.159 --> 00:41:19.920
Are we ever going to say, yeah, Claude for Legal, ChatGPT for legal, they're all the same?
00:41:20.320 --> 00:41:24.480
So I was at the Astra event and the announcement.
00:41:24.559 --> 00:41:26.400
I was literally in the front row, actually.
00:41:26.639 --> 00:41:28.960
I really liked their announcement.
00:41:29.199 --> 00:41:34.639
It is, I don't want to say it's very different, but it is different from what Claude for Legal is.
00:41:34.800 --> 00:41:38.960
What Claude did is they distilled down, as Heather mentioned, right?
00:41:39.199 --> 00:41:40.960
Patterns, usage patterns.
00:41:41.199 --> 00:41:43.440
And they they came up with the concept of skills.
00:41:43.519 --> 00:41:53.199
They came up with the concept of projects, more refined projects, and a bunch of other things, specifically spawning from their use of Claude code, from developers' use of Claude Code.
00:41:53.360 --> 00:41:56.000
So they took Claude code, they turned it into co-work.
00:41:56.239 --> 00:41:58.559
Cowork then became Claude for Legal.
00:41:58.639 --> 00:42:01.280
Um and never started out as legal, neither did GPT.
00:42:01.360 --> 00:42:08.079
But what GPT has done, they've taken a lot of what Claude did well, but also GPT's users were doing well too.
00:42:08.320 --> 00:42:13.199
They did, they got to the same conclusion, and then they added a bunch of integrations, a bunch of partnerships.
00:42:13.440 --> 00:42:16.079
That really what's mattered, what matters the most.
00:42:16.320 --> 00:42:24.880
But on top of that, what we've seen is out of the box, GPT 6 Astra is quite good at legal work.
00:42:25.039 --> 00:42:28.400
Specifically, it is quite good at not hallucinating either.
00:42:28.639 --> 00:42:38.079
And then you add in the legal integrations and plugins, you get a model that gets you high fidelity, high fidelity responses, quality responses.
00:42:38.320 --> 00:42:40.400
So I think they've done a phenomenal job.
00:42:40.559 --> 00:42:44.639
It will always be an arms race with entropic with open AI.
00:42:44.800 --> 00:42:50.000
I think one might have the edge today, another might have the edge tomorrow, but that's literally how it's going to go.
00:42:50.159 --> 00:42:56.000
I don't think that anyone, you know, you could be doing a specific job, Claude's gonna be better, or GPT is gonna be better on a given day.
00:42:56.159 --> 00:42:58.880
I think at the end of the day, the differentiation.
00:42:59.199 --> 00:43:04.480
Now, can you can an open AI or an anthropic buy a Harbor Lagora?
00:43:04.719 --> 00:43:05.039
No.
00:43:05.199 --> 00:43:07.360
The valuations are crazy, I think.
00:43:07.679 --> 00:43:11.280
But what you really then want is the workflows.
00:43:11.599 --> 00:43:13.840
Product is really easy to create today.
00:43:14.000 --> 00:43:17.599
It's being commoditized, meaning the features are being commoditized.
00:43:17.840 --> 00:43:21.760
However, there is a lot of value in those workflows.
00:43:21.920 --> 00:43:24.159
That's where companies like our ours come in.
00:43:24.239 --> 00:43:28.639
That's where maybe an Eve legal or some other of these companies come into play in question as well.
00:43:28.880 --> 00:43:31.760
So that's that's what we're seeing with the landscape right now.
00:43:31.920 --> 00:43:38.880
However, Claude and GPT have wiped out a number of AI companies, I think, and they're gonna continue doing that.
00:43:39.119 --> 00:43:44.320
That differentiation is vastly important, whether it's on the tech front and then it's gonna be the distribution and marketing.
00:43:44.480 --> 00:43:46.480
That's the those are the only things that matter today.
00:43:46.719 --> 00:43:50.559
You know, Sabi, this has been a really fascinating episode.
00:43:50.800 --> 00:43:52.559
Really enjoyed speaking with you.
00:43:52.800 --> 00:44:02.559
So if one of our listeners wants to reach out and see if it's a fit for them with uh with you and with Iris, what's their best way to do that?
00:44:02.880 --> 00:44:09.199
Either you can go on our website, fill out the contact form, we'll give you a live demo on your work, or just shoot me an email.
00:44:09.280 --> 00:44:10.079
Uh, I don't mind either.
00:44:10.239 --> 00:44:13.360
Again, we're very personable and we try to get to literally everyone.
00:44:13.519 --> 00:44:14.800
My email is subby.
00:44:15.039 --> 00:44:22.800
My first name, S-A-B as in boyi-h at ikitas.ai, I-Q I D as in David I-S.ai.
00:44:23.199 --> 00:44:24.480
That's the old name of the company.
00:44:24.639 --> 00:44:27.679
But you know, we'll we'll we'll get you sorted at the end of the day.
00:44:27.920 --> 00:44:29.360
So all right.
00:44:29.519 --> 00:44:36.400
Well, I the way I know this has been a great conversation is I didn't get to ask you any of my prepared questions.
00:44:36.639 --> 00:44:37.039
That's great.
00:44:37.360 --> 00:44:39.280
And uh that's always a good sign.
00:44:39.519 --> 00:44:40.800
So thank you so.
00:44:41.280 --> 00:44:42.400
I'm sorry, what was that?
00:44:42.800 --> 00:44:44.800
Or it means I droned on for too long.
00:44:44.880 --> 00:44:46.320
So not at all.
00:44:46.480 --> 00:44:48.639
Thank you so much for being with us.
00:44:48.880 --> 00:44:49.679
Thank you.
00:44:50.159 --> 00:44:54.559
And uh, this has been another episode of AI Tools for Practicing Lawyers.
00:44:54.719 --> 00:44:56.239
We'll see you next time.
00:44:57.280 --> 00:45:01.679
That's it for today's episode of AI Tools for Practicing Lawyers.
00:45:01.840 --> 00:45:02.960
Thanks for listening.
00:45:03.119 --> 00:45:04.719
We'll see you next time.