Latest / The Jon and Marc Podcast / Scale Without Hiring: Supio’s Blueprint for the AI-Native Law Firm

Transcript

Scale Without Hiring: Supio’s Blueprint for the AI-Native Law Firm · The Jon and Marc Podcast

• 0:00
  They're in the business of improving their models with your
• 0:02
  client data. I don't know if everybody knows
• 0:04
  that. Like I don't think anyone knows
• 0:06
  that. I can't imagine that the lawyers
• 0:08
  that are being trained today are going to be prepared for the
• 0:12
  agentic future. In this transition, a software
• 0:15
  engineer is forced to build a relationship with the
• 0:17
  technology, and attorneys need to do that as well, right?
• 0:20
  So this is probably a changing landscape.
• 0:38
  Welcome to the John and Mark podcast.
• 0:40
  We're very excited for our incredible guest today, Jerry
• 0:42
  Joe from Supio, the CEO and Co founder.
• 0:45
  If you have seen neon green advertising with shirts at all
• 0:48
  the conferences lately, that has been them.
• 0:51
  How would you define Supio correctly?
• 0:53
  I was going to say AI legal platform.
• 0:55
  I think it's. A great way to sum it up right?
• 0:57
  We're the AI partner for firms looking to jump into being an AI
• 1:02
  native firm and we're happy to help all the firms that are here
• 1:06
  and are interested in taking the dive.
• 1:08
  Awesome, quick background. So you Co founded the company
• 1:11
  with your best friend and then now you guys have grown the
• 1:15
  company to how many employees so far there's?
• 1:16
  About 130. Yeah.
• 1:18
  And then how many firms are you currently partnered with it?
• 1:20
  Over 200. So yeah, it's been, it's been
• 1:23
  quite a journey. We've been around for about four
• 1:25
  years. You know, during COVID we always
• 1:28
  wanted to build a start up and we shared our careers together.
• 1:31
  Early on. We were at a sales tax company
• 1:33
  that we helped take IPO called Avalara and there we went.
• 1:36
  Very familiar with them. Yeah.
• 1:37
  And then we went to a Microsoft for Office 365.
• 1:40
  And during COVID, we've always wanted to build a company.
• 1:42
  We, we wanted to do this in legal.
• 1:44
  And so we left and explored the market and we went and spoke to
• 1:48
  hundreds of different types of lawyers.
• 1:49
  We didn't really know that legal is so nuanced.
• 1:52
  And one day we stumbled upon personal injury and plaintiff
• 1:55
  law. And we love the fact that
• 1:57
  plaintiff lawyers help their clients in their hardest times.
• 1:59
  And also that it's always David versus Goliath story where like
• 2:03
  technology could really make a difference.
• 2:05
  And that's how we found our home.
• 2:06
  And we've kind of exploded and growth since we started.
• 2:09
  Let's talk about that timeline for a second.
• 2:10
  Four years, right? The public explosion of AI was
• 2:14
  November, I think 24. It's more like November of 23.
• 2:18
  November 23 OK, still that like burst.
• 2:20
  Onto the sea where everyone was like, have you?
• 2:22
  Used Chacha BT. So you founded the company four
• 2:24
  years ago. Was it always an AI first
• 2:27
  company? Like, did you see that coming?
• 2:29
  We saw AI was developing. I got to play with a lot of toys
• 2:32
  at Microsoft, right? And a lot of this was really
• 2:34
  around smaller language models. So Bert is an open source, one
• 2:37
  that has like a small content text.
• 2:39
  And at the time, you know, we're thinking about documents we
• 2:41
  thought that you could resolve like simple grammatical errors.
• 2:44
  If you think about building out document templates and writing a
• 2:47
  paragraph and having pronouns that like verbs and things like
• 2:50
  that, it's always a pain to actually generate documents.
• 2:53
  So, you know, we were applying those models to test that.
• 2:55
  And as these models got more complex, we realized that not
• 2:58
  only can you generate the right documents, you can use it to
• 3:00
  help understand the data to date.
• 3:02
  Today, probably the most valuable part of what we do is
• 3:05
  really helping our customers understand the mass of amount of
• 3:09
  client data that they sit on. Both it helps create better
• 3:12
  outcomes as well as just pure automation in terms of the
• 3:15
  amount of work it takes to go through every single record,
• 3:18
  aggregate these draft documents and so on.
• 3:20
  What is the end game like? You guys announced another
• 3:25
  product today, right? Prism was announced and you keep
• 3:28
  advancing what the platform actually is.
• 3:30
  What is the end game look? What's the?
• 3:31
  Vision of it, I'm super excited about the product because you
• 3:34
  guys are advancing very quickly and like where do you see it
• 3:36
  going? I love the question, how much do
• 3:38
  you guys know about agents? I know we have this buzzword all
• 3:42
  the time but like what does agentic mean?
• 3:44
  Having a fake AI person do something that a person would
• 3:48
  do. Yeah, totally.
• 3:49
  That's a great way to yeah. It's like, yeah, I should, I
• 3:54
  should take that line, I think, swing it to my team.
• 3:56
  I guess they're a real AI person, but yeah.
• 3:58
  I've built things with like Chachi PTS agents and they're
• 4:02
  not overly complex but it's telling a platform or a solution
• 4:07
  to do something and it does it for you and can be done at
• 4:10
  scale. So we look at agents very
• 4:12
  similar to this in terms of following the journey of a piece
• 4:16
  of data in a law firm. Imagine when there's a new
• 4:19
  record that comes into a paralegals inbox.
• 4:22
  The first thing they do is they have to read it.
• 4:24
  Then they have to understand what kind of document is.
• 4:26
  Then they have to put it into the right case.
• 4:27
  And then based on that, they have to understand if it tells a
• 4:30
  new story about this case that needs more evidence that needs
• 4:33
  to be ordered. And then from all of this, as it
• 4:35
  accumulates and builds on top of each other, eventually the claim
• 4:39
  is fully formed. And at that point, you're able
• 4:41
  to go generate something like a settlement package to go get the
• 4:44
  case settled, right? So that entire journey has so
• 4:47
  many different human touches and AI touches.
• 4:50
  And for us, it's about creating the full case assistant that's
• 4:54
  agentic and that's slow March into creating more structure
• 4:57
  around all these different types of data, understanding all of
• 5:00
  the rules that go along with how to build and the decisions you
• 5:04
  make. For example, if you're over
• 5:05
  treating for a client and it's going above their policy limits,
• 5:08
  then there needs to be an insight for the paralegal to
• 5:11
  say, hey, like this is probably case ready to settle.
• 5:13
  If you understand that there's missing bills or missing
• 5:16
  records, then the team needs to be able to order that on the
• 5:19
  other side and litigation, all of this diverse data that's
• 5:22
  coming in, we're helping process at scale and then we're
• 5:26
  surfacing insights that can help you make the difference between
• 5:28
  winning and losing the case as well.
• 5:30
  And all of this come together into a single agentic platform.
• 5:34
  It's an interesting way to look at it because I look at how
• 5:38
  people use AI today and they traditionally look at it as
• 5:42
  like, I'm going to give you an input and you're going to give
• 5:44
  me an output, right? I think the way that you guys
• 5:47
  are looking at it, and correct me if I'm wrong, is breaking up
• 5:49
  that process into a bunch of little small processes, right?
• 5:53
  And that's actually the right way for someone to use a
• 5:56
  ChatGPT, a Gemini, is to do like one very precise thing and do
• 6:00
  that very precise thing as well as you can do it.
• 6:03
  And then you move on to the next thing.
• 6:05
  It's a great analogy. You know, I think a lot about
• 6:07
  autonomous cars and my cruise control, the truly autonomous
• 6:10
  cars. And right now when people use
• 6:13
  something like Chachi PT, it's great.
• 6:14
  It's kind of similar to the thing kind of takes the gas
• 6:17
  pedal off. You don't have to pay as much
• 6:18
  attention. It's still driving.
• 6:20
  It's. Kind of like self driving on a
• 6:21
  highway, yes. On a highway and truly there are
• 6:24
  stretches of the road today for like autonomous Level 3 where
• 6:27
  certain cars you don't really have to pay attention and it
• 6:29
  drives by itself, right. And the moment you achieve that,
• 6:32
  that's when there's a substantial benefit to a law
• 6:35
  firm because they offload the entire process away from a
• 6:38
  people structure in the AI. So most companies haven't really
• 6:41
  reached that yet. A lot of times, you know, you
• 6:43
  offer a tool just like chat BT where you can come in and you
• 6:46
  can chat. It's an assisted workflow.
• 6:48
  Part of what Supio is trying to do is shift these assisted Volt
• 6:51
  flows. So we create a much cleaner line
• 6:53
  between where people are with the client interactions,
• 6:56
  building empathy, going and advocating versus like this
• 6:59
  document processing which now can be fully autonomous.
• 7:03
  We work with a ton of personal injury firms across the nation,
• 7:05
  and their biggest challenges as they scale is scaling the
• 7:09
  paralegals and the people that are doing that work.
• 7:12
  One is hard to find good people that want to do that work. 2.
• 7:15
  It's hard to train them in such complex processes that once they
• 7:18
  start working up a case, all the documents, all the different
• 7:21
  case types and getting those people at the speed and it's
• 7:23
  just so much manual labor. That's to be able to achieve my
• 7:26
  fairly trend. Yeah, right.
• 7:28
  Everyone's going like let me let me build a team offshore.
• 7:31
  You see vendors who are trying to build teams offshore.
• 7:33
  Is that sort of like the intermediary step?
• 7:36
  I think there's lots of things if you think about where AI is
• 7:39
  good at that AI is actually better than people at you can
• 7:41
  offshore as much as you want. In some examples, like when I
• 7:44
  think about our greatest wins, right as we've helped with a
• 7:47
  huge verdicts. A lot of times there's no amount
• 7:50
  of offshoring to go and try to digest 42,000 pages of medical
• 7:54
  records for for a case where like they've partnered the teams
• 7:58
  of nurses trying to summarize this.
• 8:00
  The summary of that is still 1000 pages.
• 8:03
  And now you try to go understand that with Supio, because we
• 8:05
  break down all that data into something digestible.
• 8:07
  You can actually take the product to court.
• 8:09
  And when someone's asking about the interaction between the
• 8:11
  nutritionist and the baby for like any C case, you can
• 8:15
  retrieve it at the moment, right?
• 8:17
  And that's the power of AI. It actually up levels the
• 8:19
  standard of certain things that as many people as trained as
• 8:22
  possible. There's a limitation there.
• 8:24
  Amazing. You guys did have part in a very
• 8:28
  large win. You want to tell everybody about
• 8:30
  that if they haven't heard about it.
• 8:32
  Yeah. It was last year we were first
• 8:34
  introduced to this amazing firm called Tor Herman, and we were
• 8:38
  partnering with them on a bellwether case for necrotizing
• 8:41
  enterocolitis. And this is against Abelabs.
• 8:44
  At the time. It was such a huge data
• 8:46
  challenge for us because they came to us with all of this data
• 8:49
  that is hard to digest, right? And it's so dense.
• 8:52
  It's like 10s of thousands of pages of records and part of
• 8:56
  what got us really excited, invited is the fact that AI can
• 8:59
  take you to the next level and understanding this, that's
• 9:01
  everything from the contextualization of just having
• 9:04
  a conversation with this data and being able to fetch the
• 9:07
  answers really accurately to being able to recall a
• 9:10
  chronologies that's completely interactive that you can drill
• 9:12
  into, especially with more records and in the dynamic
• 9:15
  setting of court. We understood the state and
• 9:17
  helped our firms and customers understand data that ultimately
• 9:21
  provides better outcomes for their clients.
• 9:24
  What are some of the things that you think will never make it to
• 9:28
  being transitioned to area? Yeah, right.
• 9:31
  A human has to do it. I think that the legal world,
• 9:34
  the foundation of this is to serve people, right?
• 9:37
  So this human interaction between like an attorney and
• 9:40
  building that relationship and building trust with a client for
• 9:44
  like what is the settlement to take?
• 9:46
  I imagine like I listened to stories from our customers and
• 9:48
  every single time an insurance company comes back with some
• 9:51
  type of settlement, it's so stressful for the client, right?
• 9:54
  Do I take it, do I not? Do I go and go to court?
• 9:56
  And that type of advocacy and being an advisor to the client
• 10:00
  is a very human act that will never go away.
• 10:03
  What AI can do to make this better is arm the lawyer with
• 10:07
  all the data to help the client make a better decision, and I
• 10:10
  think that's going to be the part that's going to be the game
• 10:12
  changer. Law school has prepared
• 10:16
  attorneys for the things that Supio is now doing right Briefs
• 10:22
  and drafting and demands. I don't know, I'm not an
• 10:25
  attorney, but I can't imagine that the lawyers are being
• 10:29
  trained today in law school are going to be prepared for the
• 10:34
  agentic future. What's your perception of of how
• 10:37
  that is going to play? Out.
• 10:38
  Well, that isn't just lawyers. As far as like most people going
• 10:41
  through college, sure. But but in the legal space,
• 10:43
  right, like this is you're talking about like the thing
• 10:46
  that cannot be done by AI is this human interaction.
• 10:50
  And when you're in law school, you're like you're studying
• 10:52
  books, you're writing papers, you're not human interacting.
• 10:56
  How do you see that transition happening?
• 10:58
  A great analogy. You know, I'm a software
• 11:00
  engineer by trade. So today probably the most
• 11:03
  advanced field in AI applications is in software
• 11:07
  engineering where, you know, you read about on the news like
• 11:09
  cloud code, all these companies that are generating code.
• 11:12
  So it's really changing the landscape of what engineers can
• 11:15
  do in reality, that change, which means that there's a lot
• 11:18
  less people drafting, writing code, which is like drafting
• 11:21
  documents, and a lot more people making higher level decisions.
• 11:24
  But no matter what, in this transition, a software engineer
• 11:27
  is forced to make a build a relationship with the technology
• 11:30
  and attorneys need to do that as well, right?
• 11:32
  So this is probably a changing landscape right now.
• 11:36
  Briefs and demand letters and all of this work, they are what
• 11:40
  we call artifacts that you generate, but you should have an
• 11:43
  understanding of what it's actually going to do before it
• 11:46
  writes it. So the more you use a product
• 11:48
  like Supio, the more predictable you understand how it's going to
• 11:51
  behave. And that builds a relationship
• 11:53
  not unlike learning how to use the Internet or your iPhone or
• 11:57
  any of these other pieces of technology.
• 11:58
  What have you seen as the biggest challenge for getting
• 12:01
  firms to understand this or to adopt this?
• 12:03
  They said, hey, we want to move forward with you, but like
• 12:06
  implementing. What's the biggest hurdle you've
• 12:07
  seen them adopting this technology?
• 12:09
  It's like every person in the world, there's a lot of
• 12:12
  resistance when people just don't understand what you're
• 12:16
  trying to force upon them. Yeah.
• 12:18
  Exactly. Software implementation is
• 12:20
  actually painful. And AI is actually different in
• 12:22
  my experience. It's like as a software
• 12:24
  implementation, this is one of the rare pieces of software that
• 12:27
  actually helps you with your case.
• 12:29
  But if you think about the dimensions of like practice
• 12:31
  management system to all the software of the past, you know,
• 12:33
  that's really like a very administrative piece of
• 12:36
  software. It's like additional work.
• 12:37
  For you? For you data entry tracking.
• 12:41
  So a lot of this is making sure on the the firm to the employee,
• 12:45
  this is a piece of software that is supposed to help you with the
• 12:48
  way you work on the case. So I think the aha moment for a
• 12:51
  lot of people is just being able to interact with the software
• 12:55
  and asking the question and understanding like, Oh my gosh,
• 12:58
  I can contextualize my questions.
• 13:00
  I can follow up with lots of streams of theories about this
• 13:03
  case. I can have a conversation with a
• 13:05
  client and instantly get an answer back without spending
• 13:08
  days to get my pair legal to go and summarize something for me,
• 13:11
  and that's a game changer in this field.
• 13:13
  What? Challenges have you had
• 13:15
  implementing this across different firms because you guys
• 13:18
  are kind of almost agnostic because they all have different
• 13:21
  case management systems, they all have different operations
• 13:23
  and how they do things. What's that been like?
• 13:26
  It's been challenging. The primary thing that's helped
• 13:28
  us navigate all of this is we're a very data first company,
• 13:32
  right? Even when we started the
• 13:33
  company, we started meeting personal injury lawyers and just
• 13:36
  listening to them and looking at their cases.
• 13:39
  We realize that all the data flowing into a personal laundry
• 13:42
  case is very similar. And if you can break down that
• 13:45
  data, all of these things are universal across all these cases
• 13:48
  and the work you do to process them, even if the outputs are
• 13:50
  slightly different, it's the same degree of understanding and
• 13:53
  the relationships you're trying to tie this data together.
• 13:55
  So we follow that data, we ingest all of it, and then from
• 13:58
  there we help firms build out the workflows that they want on
• 14:01
  top of that. There is no standardized data
• 14:03
  model. So law firms, yes, when we
• 14:05
  worked at Litify. Like, even firms that use the
• 14:07
  same practice management system each have their own data model
• 14:11
  because they all have unique processes.
• 14:12
  Yes, totally. You're, you're 100% right.
• 14:16
  And, and the difference between those types of data models is
• 14:19
  they're looking at the outputs, the work that's trying to be
• 14:21
  accomplished. We look at the ingestion, the
• 14:23
  documents that are coming in. It can get as nitty gritty as
• 14:26
  like this is a medical record from an Epic healthcare
• 14:29
  hospital. And like, there's a nursing note
• 14:32
  that's diagnosing something like a kidney cancer for PFAS, right?
• 14:36
  And for a nursing note, well, that's not a precise diagnosis.
• 14:39
  There's also a clinical diagnosis where, you know, the
• 14:42
  doctor actually breaks the news to the patient.
• 14:43
  There's a pathology report. There's an actual like suspected
• 14:47
  diagnosis for this type of cancer as well, right?
• 14:50
  So it gets into the nuances of understanding this data to the
• 14:53
  next degree that replicates what humans are doing today.
• 14:57
  So much information. Tell me about it or Kim says
• 15:01
  that all the time This, this is what we, we've dedicated our
• 15:04
  careers to figuring out, right? And that's why all these
• 15:07
  releases come out. It's an endless deeper
• 15:09
  understanding of this data to kind of service our customers
• 15:12
  this way. How do you keep PII secure when
• 15:18
  it's going through an AI platform?
• 15:20
  Because that's like the biggest fear of lawyers.
• 15:23
  Like Oh my gosh, what happens if every file that I upload winds
• 15:26
  up somewhere else on the Internet?
• 15:28
  That this is table stakes today and incredibly important, right?
• 15:31
  The first thing is that you have to have zero day retention
• 15:35
  policies with foundation model providers.
• 15:37
  So folks like Open AI and Anthropic, right?
• 15:39
  They're not allowed to actually keep any logs of the data that
• 15:42
  we send over to them. The second one is HIPAA and sock
• 15:45
  2 compliance. We're sock 2 type 2 compliant,
• 15:47
  which means that we're constantly being audited and we
• 15:49
  offer all sorts of authentication protocols and
• 15:51
  logs to understand how things are accessed as well.
• 15:54
  So authentication really basically like 2 factor auth
• 15:57
  internally for our own teams and externally offered for firms
• 16:01
  that are concerned about their security, right?
• 16:03
  And just constantly locking this down and up.
• 16:05
  Levelling our game is how we keep our client data secure.
• 16:08
  I want to be clear, right? For those of you watching like,
• 16:10
  that's very different than someone who's using ChatGPT to
• 16:14
  write a demand letter, how does Open AI handle the similar data
• 16:18
  that would someone would upload through Supio versus uploading
• 16:21
  just through standard ChatGPT 5? For companies like that with
• 16:24
  something like ChatGPT or Gemini or Claude, right, what they're
• 16:29
  trying to do is get that data to continue to reinforce and refine
• 16:32
  their models. That's their business model.
• 16:34
  And then they take that fine model and then they sell it to
• 16:36
  enterprise companies like ours that actually need secure data,
• 16:39
  right? So you probably don't want to be
• 16:41
  uploading your data into these types of portals because they're
• 16:44
  they're in the business of improving their models with your
• 16:46
  client data. I don't know if everybody knows
• 16:49
  that. Like, I don't think anyone knows
• 16:51
  that. No, we have a.
• 16:52
  We have a policy on our team like you cannot use any AI tool
• 16:56
  unless it has been pre approved and even if it has been pre
• 16:58
  approved, you're never uploading any client data, you're never
• 17:01
  uploading any lead data, you're never uploading any Pai, you're
• 17:04
  never uploading any financial data.
• 17:06
  We had to lay those ground rules because things move so quickly.
• 17:09
  Open AI released their browser today, right?
• 17:11
  Like if people want to start using it and I'm like, no, you
• 17:14
  cannot yet. Like we have to set what the
• 17:17
  rules are around these things so that the data that's going in
• 17:20
  them is secure. And it's good to know that you
• 17:22
  guys have these proper protections in place because you
• 17:24
  are dealing with 200 plus different law firms and all
• 17:27
  their client data. You're totally right.
• 17:29
  I mean, the reality is that a lot of these firms with the
• 17:32
  staff, they're not necessarily educated on data security.
• 17:35
  So by being a vendor there, we help provide that level of
• 17:38
  education. And it just creates a level of
• 17:40
  control that incentivizes from owners to make these decisions
• 17:43
  quicker. So you can put these systems in
• 17:45
  and be ready for the future as these tools kind of grow.
• 17:48
  What are some of the most common objections you hear when you're
• 17:51
  talking to new firms about adapting the?
• 17:53
  Platform. Probably the biggest one is that
• 17:55
  they just want to make sure that their staff is ready for they've
• 17:58
  all been through these like terrible rollouts of software
• 18:01
  where it's taken a year. I know you guys know what we're
• 18:04
  talking about and they're just worried that their staff like
• 18:07
  they don't want to shove more software down their throat.
• 18:09
  They're already asking the staff to be heroes at everything,
• 18:12
  right? And you know, for us, we we
• 18:15
  enter these relationships telling them that we'll be a
• 18:17
  partner not only to them, but their staff, like we should be
• 18:20
  helping up level their game, sharing with them the best
• 18:22
  practices around AI. And this is not just a about
• 18:25
  building or learning how to use our product, but building a
• 18:28
  relationship with AI in general for the team.
• 18:31
  And as we get into the work they do, whether it's reviewing
• 18:34
  medical records, drafting demands to actually going into
• 18:37
  litigation, we bring our solutions consultants who are
• 18:40
  all licensed attorneys in the United States.
• 18:42
  They know what they're talking about and they can kind of go
• 18:44
  and interact on a deeper professional level with these
• 18:47
  teams to make them feel more comfortable as we roll this out.
• 18:51
  That's great. I didn't realize all your
• 18:53
  solutions consultants were licensed attorneys.
• 18:55
  They know the job, they know the processes, they can speak the
• 18:57
  language instead of someone who knows the technology who doesn't
• 19:00
  really know legal coming in there trying to implement it and
• 19:03
  force it down someone's throat. Usually a disconnect.
• 19:06
  Yes, totally. Five years from now, where do
• 19:09
  you think the technology is going to be?
• 19:12
  I think we'll reach. I'll even give you like 18
• 19:15
  months. You choose.
• 19:18
  You choose your time frame. Like 36 months is a good, good
• 19:21
  number, right? And I like, I think about
• 19:24
  technology from the lens of how many different types of outcomes
• 19:29
  or circumstances there are for a given task.
• 19:32
  So if you get into litigation, for example, the number of
• 19:35
  outcomes like explodes exponentially depending on the
• 19:38
  data and the jurisdiction and the the judge you have.
• 19:41
  So it's so nuanced that we're probably not going to remove any
• 19:46
  type of like attorney interaction from there anytime
• 19:48
  soon, right? But if you get into something
• 19:50
  like pre litigation or intake or these like very routine things
• 19:54
  where you know what the work is, you're trying to stand it, then
• 19:57
  you're going to be able to go and up level your staff to focus
• 20:00
  a lot more on the client. Because the heavy lifting of the
• 20:03
  actual operations can all be done where you can monitor call
• 20:07
  quality, not be able to miss any major leads.
• 20:10
  Being able to transfer order all the records, prepare all the
• 20:13
  data, give you insights so you can better coach your clients,
• 20:16
  right. All of that type of
• 20:17
  understanding that people spend with their data will be all
• 20:20
  automated. Amazing.
• 20:21
  I've been harping on a bunch of episodes.
• 20:23
  I believe the major differentiator for for companies
• 20:26
  in the future and maybe in the next 18 to 24 months, or the
• 20:29
  companies that have really good personal connections with their
• 20:32
  clients. Everyone's trying to automate
• 20:33
  everything. But like if you can maintain a
• 20:36
  really good personal connection where you can talk to someone on
• 20:38
  the phone and answer your question, but you're able to
• 20:40
  automate all the other stuff. Those are the companies that are
• 20:42
  going to be successful versus the ones that try to automate
• 20:45
  everything. And it's impossible to deal with
• 20:47
  because you're just dealing with IVRS and agents and you just
• 20:50
  want to talk to a real person. But they're able to get all the
• 20:52
  other stuff out of the way and automate it and have great
• 20:56
  humans interacting with their clients.
• 20:57
  Yeah, that's an interesting concept to like think about,
• 21:00
  right? Instead of using AI to increase
• 21:03
  your profit margins, you're talking about using AI to free
• 21:07
  up time so that you can have more personal interactions, you
• 21:11
  can have better client outcomes. You're using it as a path to
• 21:15
  really the same end goal, not necessarily just squeezing out
• 21:19
  more profit. Yep.
• 21:21
  More profit is ideal as well, but yes.
• 21:24
  Like that comes in, I think that comes along with it, right?
• 21:26
  If you are providing more human interactions, if you are having
• 21:30
  better client outcomes, inevitably you will have better
• 21:32
  profit margins. If you use all AI and no human
• 21:35
  interactions, you may have short term profit gains but then long
• 21:38
  term things dry up. At the end of the day, I think
• 21:42
  law firms, like other businesses, they AI gives them
• 21:46
  time to free up their space to think about what differentiates
• 21:50
  them, like what makes law firm special compared to their
• 21:53
  competitors. And a lot of the time, I think
• 21:55
  if you really drill into it, it is their client service, their
• 21:58
  advocacy, their ability to understand when to settle and
• 22:02
  when to fight. And that just gets highlighted
• 22:04
  as the other parts are table stakes.
• 22:09
  Very well said. This has been a fantastic
• 22:12
  conversation. This is my favorite souvenir
• 22:15
  coming from MTMP. This is the Supio gambling chip.
• 22:19
  I was told by the person who gave it to me that it is also
• 22:22
  the color of $1000 chip at the Bellagio.
• 22:24
  So no pressure do with that knowledge as you as you so
• 22:28
  choose. Fantastic conversation Jerry,
• 22:30
  really appreciate you sitting down with us.
• 22:32
  Enjoy the rest of the conference and we are very excited to see
• 22:35
  what's on the horizon for Supio. Thank you so much.
• 22:37
  We could Chad, you know, for another hour.
• 22:39
  But definitely I. Appreciate your time.
• 22:41
  Well, thanks for having me guys. Appreciate it.