Latest / Technically Sound Strategy / Episode 40 - From Concept to Practice: Operationalizing New Tech in Law Firms With Pete Terryn

Transcript

Episode 40 - From Concept to Practice: Operationalizing New Tech in Law Firms With Pete Terryn · Technically Sound Strategy

• 0:01
  Welcome to Legal Marketing Radio, a production of LaFleur Marketing.
• 0:05
  This week, I chat with Pete Terryn, Director of Operations
• 0:07
  for New Wave Technology Partners.
• 0:10
  Pete and I spent some time talking about operationalizing new tech in law firms.
• 0:14
  In addition to his work with New Wave Technology Partners, Pete is passionate
• 0:18
  about artificial intelligence.
• 0:20
  He co founded the West Michigan AI Lab, a vibrant community hub dedicated
• 0:24
  to connecting AI enthusiasts with valuable insights and resources.
• 0:29
  He also brings AI expertise to a broader audience as co host of the
• 0:33
  popular podcast, AI Lab Unfiltered, where he engages expert guests and
• 0:38
  deep dives into their specialties.
• 0:41
  Beyond his professional endeavors, Pete is actively involved in his
• 0:44
  community as a high school basketball coach in Caledonia, Michigan.
• 0:47
  He balances his professional and community commitments with his role as a devoted
• 0:51
  husband and father of three teenagers.
• 0:54
  All right, Pete.
• 0:55
  Really glad to have you here with us today.
• 0:56
  So In our conversation leading up to this, we covered kind of a range of topics,
• 1:01
  but, in light of our audience and some of the things that I keep hearing about
• 1:06
  in this space, as I talk to lawyers, all over the country is really the
• 1:11
  operationalization of AI across the firm.
• 1:15
  And I think we're all in a growing evolutionary stage of that.
• 1:20
  And we use AI across our organization in a lot of different ways, but
• 1:24
  then the operationalization, of it is a separate thing.
• 1:27
  And so I know you've got a lot of experience in that space.
• 1:30
  So I guess, give me a little background help me understand how you, what
• 1:34
  your relationship with AI looks like and how you used it within your
• 1:38
  current role, if you would, please.
• 1:40
  Yeah.
• 1:41
  Thanks for having me.
• 1:42
  It's my favorite thing to talk about.
• 1:43
  So I'm glad to have an opportunity to do that.
• 1:46
  My own background, I'm not one of these people that's going to say that
• 1:49
  they started in machine learning, 20 years ago or anything like that.
• 1:53
  I'm Johnny come lately on the AI scene.
• 1:56
  I didn't really pay much attention to it until, GPT three came out.
• 2:00
  Just , two and a half years or so that I've been since then, though,
• 2:04
  really every day I'm reading about it.
• 2:06
  I've got multiple emails that come into my inbox that have daily digests of
• 2:11
  the news and what's happening with AI.
• 2:13
  It's kind of, I call it my soap opera, all the drama with all the
• 2:16
  big players and what they're doing.
• 2:19
  And then I'm, spending time on Reddit researching, different
• 2:22
  technologies that are.
• 2:24
  The different advancements with AI and how people are using it.
• 2:27
  I'm also been fortunate enough to speak in front of a bunch of businesses about
• 2:31
  AI, the Small Business Association of Michigan here, World Trade Week last fall.
• 2:37
  And also with another gentleman named Drew Veach who works with
• 2:41
  a company called BizStream here started a monthly AI meetup.
• 2:46
  Last year, last January, we've been doing that for over a year now, where
• 2:49
  we bring in featured speakers and folks from around the area to learn about AI.
• 2:54
  And we just launched our own podcast called AI Lab Unfiltered.
• 2:58
  I'm in it, I'm around it, I'm immersing myself in it.
• 3:02
  I'm just trying to meet as many people as I can and trying to build
• 3:05
  a really strong AI ecosystem here.
• 3:08
  As far as my background, I'm not a technical expert.
• 3:11
  I'm not, Writing AI agents or anything like that.
• 3:15
  I'm not a coder, but I love this stuff and I love talking about it and I love
• 3:19
  helping people become familiar with it and get over that fear of the unknown
• 3:23
  that's holding a lot of people back.
• 3:26
  I think that's the case operationally to get to the main topic here is,
• 3:32
  there is A lot of fear around it.
• 3:34
  And one of the fears is that I'm already so far behind that I can't start with AI.
• 3:39
  My, my competition's already so far ahead of me.
• 3:42
  I'm sure they're in some, dark room coding away and they're going to
• 3:46
  automate everything, drop their prices and take me out of the market.
• 3:50
  And the reality is, and I just talked, I had coffee with a gentleman this
• 3:55
  morning who went to South by Southwest.
• 3:57
  Last week and was hanging around a bunch of people in AI and one of the things that
• 4:01
  stood out to him and he came back and he said, nobody's hardly anywhere with this.
• 4:06
  There's very few companies that are doing anything with AI and
• 4:09
  everybody else thinks that everybody else is doing something with it.
• 4:13
  So it's not too late to start.
• 4:15
  This is a great time to start.
• 4:17
  You don't have to go back and do two and a half years of research to get caught up
• 4:21
  because the technology changes so quickly.
• 4:24
  That if you just started learning now, within a couple of weeks or a month, if
• 4:29
  you were to do just invest a half an hour to an hour a day, which is what I'm doing.
• 4:34
  You'd be caught up.
• 4:35
  Pretty darn good in that time.
• 4:38
  So don't let that fear hold you back.
• 4:40
  If you're thinking about implementing AI from an operation standpoint and
• 4:45
  the other one is that this is only for geniuses , to take on and you have to have
• 4:50
  a PhD in machine learning or artificial intelligence , to implement this.
• 4:55
  I don't have a college degree, so I can tell you that's not the case,
• 4:58
  and we've implemented some AI here at New Wave, so it's more about having,
• 5:04
  I think that's the first big key for me from an operations standpoint,
• 5:07
  is you have to have an AI champion.
• 5:09
  You have to have someone that's going to lead the charge.
• 5:13
  I just happen to be in a leadership role here, which helps.
• 5:17
  But you don't have to be, you don't have to have a director
• 5:20
  title or a c in your title , to implement AI in your organization.
• 5:24
  You should be connecting with those folks and sharing your passion for it and
• 5:28
  that you are knowledgeable in that area.
• 5:30
  And if they're smart, they'll use you as a way , to help implement it.
• 5:33
  So I think.
• 5:34
  The big thing really is identifying a champion first, and if you don't have one,
• 5:39
  maybe you need to hire one because if you don't, I think your chances of getting
• 5:43
  something implemented are pretty low so you might need to go look for somebody
• 5:47
  that's doing that or find a peer that you have a great relationship with learn what
• 5:52
  they're doing, if they're doing anything with AI, and this Join our meetup.
• 5:57
  If you're in West Michigan you'll bump into people and you'll learn who's doing
• 6:01
  what and how you can learn from them.
• 6:02
  But there's ways to do it.
• 6:04
  But starting, having a champion, I think, is that first one.
• 6:08
  Yeah,
• 6:10
  that makes sense to me.
• 6:11
  You covered some ground there.
• 6:12
  But one of the things that jumped out is the computer scientists, the
• 6:17
  data engineers, the folks that have been working in machine learning
• 6:20
  for years and years, like the change that we have experienced is not
• 6:25
  something that requires that level of education, that level of background.
• 6:29
  And I think one of the biggest misconceptions that I've run across is
• 6:33
  this idea that AI is brand new and that, all of these tools just appeared on the
• 6:39
  scene out of somewhere and they showed up.
• 6:42
  And the reality is AI has been around for quite some time and some of the
• 6:46
  functionality and some of the use cases that have been really cool.
• 6:50
  Have lived in that developer space and in that highly technical space
• 6:55
  and just haven't been able to bridge the gap and they haven't been able
• 6:58
  to bridge the gap because I think that organizational leadership was
• 7:03
  not prioritizing the implementation of that technology a few years ago.
• 7:08
  We built a tool that would basically.
• 7:11
  Pull in a bunch of case data from a trial attorney down in Louisiana.
• 7:15
  They gave us about 40, 000 records to train our model.
• 7:18
  We built a model.
• 7:18
  The model could then make some level of prediction if a case was
• 7:21
  going to be a paying case or not.
• 7:22
  So meaning if a plaintiff would come to the table with injuries
• 7:27
  with the correct coverage and the correct party at fault, right?
• 7:31
  And so we could analyze that and figure that out pretty quickly.
• 7:34
  But we didn't find it that much appetite for it because all the lawyers that I
• 7:38
  talked to didn't think it was possible and we'd show them, and they still
• 7:42
  just didn't, they didn't have buy in.
• 7:44
  And so I think the thing that has changed is that people like you and I hope me
• 7:49
  have looked at this and seen that the new use cases and how the companies
• 7:54
  behind it, I think have made it so accessible that it becomes available.
• 7:59
  But, to your point, like , it hasn't existed in the boardroom and I think
• 8:03
  that's just starting to happen now.
• 8:05
  I think the approach makes sense, having a champion, who's going to come in there.
• 8:09
  So if you don't mind, give me some examples of.
• 8:11
  How you've seen a successful implementation.
• 8:13
  Give me some like use cases where you've seen, where you've seen it work.
• 8:18
  Where the AI is actually doing work within the organization.
• 8:23
  Yeah, yeah, where you've built some processes, you've built some SOPs
• 8:26
  or something along those lines.
• 8:29
  Yeah one thing that we built that's another key to this is that it built
• 8:34
  its own AI policy for the organization.
• 8:36
  An acceptable use policy.
• 8:39
  We met as a leadership group and we determined pretty early on that if we
• 8:42
  were going to, and we're an IT firm, so we are concerned about things like
• 8:46
  data security and things like that.
• 8:49
  And we were starting to talk about AI in our quarterly all staff meetings
• 8:53
  and just me just trying to get people excited about it and ready to leverage it.
• 8:57
  And we said now people are going to start using it.
• 8:59
  What are they going to use it for?
• 9:01
  And how do we get ahead of that?
• 9:02
  So we created a policy that governs the proper use of it.
• 9:07
  And some of those, some of that policy is like the actual
• 9:12
  work that is generated with AI.
• 9:14
  If you're submitting it, that's your work.
• 9:18
  You can't if something goes wrong or you didn't research to make sure that
• 9:21
  it was factual and you present wrong information, you can't just go back
• 9:25
  and say, well, that's AI's fault.
• 9:27
  Well, no, that's, we want to be very clear that was your work.
• 9:30
  So for those that are listening and don't have an AI acceptable use
• 9:34
  policy, or they have one that doesn't have that in it, you can take that
• 9:38
  as a way to, to implement that.
• 9:40
  We still took that document and sent it to our lawyer for review to make
• 9:44
  sure that, it was a good document.
• 9:47
  So we still, it have that piece to it . But the drafting of it
• 9:50
  was more within our control.
• 9:52
  It more reflected our thoughts as a leadership team and
• 9:57
  what we were doing with it.
• 9:58
  We've also implemented AI as an augmentation to our technical staff.
• 10:04
  So we are a help desk for companies that don't have their own help desk.
• 10:09
  So solving PC issues and internet goes out or, security, things like that.
• 10:17
  Where our technicians are referencing using an AI tool.
• 10:22
  That we've built prompts around to help get to a faster determination
• 10:27
  of the technical issue for that client, and it helps categorize
• 10:32
  those tickets into a way that we can look them up easier in the past.
• 10:37
  And just.
• 10:38
  And this is, I think, the greatest and best use of AI.
• 10:41
  We talked about replacing people.
• 10:43
  No, I think the greatest and best use of AI is augmenting people.
• 10:47
  Giving them superhuman powers that they wouldn't have otherwise.
• 10:51
  So you're leveling everybody up.
• 10:53
  So you have a level one technician who comes in new to the industry.
• 10:57
  They've only been doing it for, six months or a year.
• 10:59
  They've got this second brain.
• 11:01
  now that can help get them to a resolution faster.
• 11:06
  Yeah, it makes sense.
• 11:07
  What are you doing in terms of controls?
• 11:09
  Like you mentioned you mentioned making sure that people realize and
• 11:13
  that people buy into the idea that your work is still your work, right?
• 11:17
  We actually just had an internal kind of lunch and learn, I think it
• 11:20
  was a brunch and learn technically to talk about copyright, right?
• 11:24
  What's trademarkable?
• 11:25
  What's copyrightable?
• 11:25
  What's important in our space?
• 11:27
  But we touched on a little bit of that, but what are some controls that you put
• 11:30
  in place, apart from just, that clear accountability, which I do think is huge.
• 11:34
  I don't want to underscore the importance of that, but what have
• 11:37
  you built out in terms of controls to but by department by area of focus.
• 11:42
  So we're not really producing anything that's externally.
• 11:46
  Facing aside from maybe some of our social media posts and some content there and
• 11:52
  we're not basing it off of something else.
• 11:55
  it's generative, it's based off of a topic, not necessarily a
• 11:58
  document where we say, hey, make something like this , we are mindful.
• 12:03
  Now, do we have.
• 12:04
  a policy that says don't do that and be lying if I said we did, but we've had
• 12:10
  enough of those conversations to say.
• 12:13
  We're not reproducing other people's work there
• 12:16
  More along the lines of, the IT space.
• 12:19
  So if someone writes some code, apart from just the ownership, right?
• 12:24
  Someone writes code.
• 12:25
  Have you modified your testing processes?
• 12:27
  , and I asked in line with, within a law firm, there are controls and mechanisms
• 12:32
  that need to be in place, right?
• 12:33
  So , one of the things that will happen is they'll take out a client,
• 12:36
  they'll get all the documents from the client and they'll write a demand
• 12:39
  letter and the demand letter says.
• 12:41
  Because of these circumstances, because this person was at fault,
• 12:44
  because they have the coverage, because of these injuries, we've
• 12:47
  documented those the liability ends up being X amount of dollars.
• 12:53
  Now, I could see a scenario where you enter all those
• 12:55
  pieces of information, right?
• 12:57
  Accident report, medical records.
• 12:59
  They enter that in and write a demand letter, but I could very
• 13:03
  easily see someone in a firm.
• 13:05
  Doing that using off the shelf chat GPT and then lo and behold, that data is being
• 13:12
  used to train the model on how somebody's , HIPAA protected, personally identifiable
• 13:16
  information is out there at least.
• 13:18
  Now you guys are dealing with that, but you are dealing with technical execution.
• 13:20
  Things can go wrong.
• 13:21
  So I guess, can you expound on the controls?
• 13:24
  From that perspective.
• 13:25
  Yeah, thank you for clarifying.
• 13:27
  So that is your data of training the model your data, especially proprietary data,
• 13:34
  leaving your environment is a big concern when it comes to using these AI tools.
• 13:40
  In my experience, when I'm speaking with different groups, I'll recommend 1 of
• 13:44
  2 platforms that have been very vocal about, at minimum, giving you a toggle
• 13:49
  to say, don't train the data with my information, that'd be Microsoft and what
• 13:54
  they're doing with Co Pilot and OpenAI.
• 13:57
  You can.
• 13:58
  Now, whether or not they're doing what they're saying that they're
• 14:01
  doing , I don't work for either of those companies, so I'm not representing
• 14:04
  them, but if you read all the disclaimers, they're pretty adamant.
• 14:08
  And I have a friend who's a Microsoft AI MVP, and talk with him, and
• 14:13
  he works with Microsoft directly, and he feels pretty confident.
• 14:16
  his kids.
• 14:17
  Use is Microsoft AI in, the Azure EN environment.
• 14:21
  So if he trusts it I trust him and so , I trust it.
• 14:25
  Same with open ai, they're pretty clear about, you can flip that toggle and
• 14:28
  say don't train the model of my data.
• 14:31
  not to say that other platforms don't, but I haven't gone and researched all of
• 14:34
  them because those are the two that I use the most, but I think if you're just do
• 14:39
  your due diligence, whether you're going to a direct model, like open AI or a
• 14:45
  perplexity or Google Gemini, look for that and make sure that there's something that
• 14:51
  blatantly says we're not using your data.
• 14:54
  Now, what you could Also run into is you're buying a product that is what's
• 14:59
  called a wrapper around one of those models like a chat GPT and a lot of stuff
• 15:06
  like Harvey is a tool that law firms use.
• 15:09
  It's a wrapper that on chat GPT, that's the base model.
• 15:13
  So you would want to talk with that company to develop that wrapper and ask
• 15:18
  them to in writing what they're doing with that data to make sure that they're not
• 15:23
  allowing that data to train the model.
• 15:27
  And if they are, then it's up to you whether or not you want to
• 15:29
  feed that data into the model.
• 15:32
  Yeah, it sounds like AI is going to be the cause and the solution
• 15:35
  to many legal headaches, even within legal clients, right?
• 15:39
  For all of those reasons, we, I I didn't, fortunately I didn't make a big mistake,
• 15:43
  I've made the mistake early on talking about Chad GPT, and their pro level is
• 15:48
  being like one of the levels that doesn't use your model to train and you are
• 15:51
  right, like you have to shut that off.
• 15:55
  And it's not until you get to the team's version of chat GPT where
• 15:58
  it comes by default turned off,
• 16:00
  but that
• 16:01
  is not necessarily intuitive either.
• 16:03
  And I don't think that the process of shutting it off in
• 16:06
  the pro model is intuitive.
• 16:07
  So , I've been telling the clients that we work with that are more
• 16:11
  excited about AI that are more interested in experimenting, which I
• 16:13
  think is a really important part of.
• 16:15
  Operationalizing to just go to the teams, chat, GPT teams.
• 16:19
  I agree with Charlie's perspective though.
• 16:21
  If you can get into Azure, if you can get into open AI through Azure,
• 16:25
  I think that's a great way to do it.
• 16:26
  I think then you have a commitment from Microsoft and you have
• 16:29
  a commitment from open AI.
• 16:30
  So it seems like that's the best way to do it.
• 16:32
  I've not played too much within Azure because it seems like it's much more
• 16:36
  of a developer platform I've used copilot a little bit, so I guess,
• 16:41
  my next question and you tell me what you think, but which platforms
• 16:46
  do you feel excel in which areas?
• 16:49
  What are your go tos in your day to day?
• 16:52
  For a while the perplexity platform, which if folks aren't familiar with it, it was
• 16:57
  really, In a league of its own when it first came out because it was doing what
• 17:02
  other models are calling deep research.
• 17:04
  Yeah.
• 17:04
  Right now is where it goes and it does its own kind of self-directed research.
• 17:09
  I don't know how nerdy , you want to get with this, but, um, super nerdy.
• 17:13
  So nerdy.
• 17:14
  Go all the way, man.
• 17:15
  All all the way into nerdy land.
• 17:17
  . Alright,
• 17:18
  so I, there, they're still very strong in that area, but the what's this whole
• 17:24
  cat and mouse game with these frontier labs is as soon as they see somebody
• 17:29
  else doing something cool, it's only a matter of months before, sometimes weeks
• 17:33
  before they're doing it themselves.
• 17:35
  Yeah, so I would still say they're the clubhouse leader there because
• 17:40
  they've been doing it longer.
• 17:42
  A great platform.
• 17:43
  Perplexity.
• 17:44
  When you're when you want to do something From a real research standpoint.
• 17:47
  And so what it'll do is you'll come to it with a prompt.
• 17:51
  And you'll say, I wanted, I just did this a couple weeks ago.
• 17:55
  I wanted to get, March Madness was coming up and I wanted to have a
• 17:58
  new TV that was great for sports.
• 18:00
  So I had a great, refresh rate and the whole thing.
• 18:03
  So I asked it to, research the best TV to watch sports with.
• 18:07
  And it took, I don't know, five, 10 minutes of, and it came back with.
• 18:12
  a whole breakdown of the different TVs.
• 18:16
  And the cool thing before it even started was that it asked me a couple of questions
• 18:20
  like, what size are you looking for?
• 18:21
  And is there a specific brand that you prefer?
• 18:25
  And I said, Nope, just go find , the best TV.
• 18:27
  And it came back with its own research, probably six or eight different
• 18:30
  TVs, links to where I could buy them on Amazon and the whole thing.
• 18:34
  So if you're doing that kind of research, I think Perplexity is still
• 18:37
  the leader, but Gemini can do that now.
• 18:39
  OpenAI has that now with the different chat GPT models.
• 18:43
  For those of you who are coding inclined, you like to write code, Claude is
• 18:48
  the place to go for coding right now.
• 18:51
  I'm not a coder, I'm not a developer, but I've read enough
• 18:55
  and learned enough about it to know that's the place to go for that.
• 18:58
  OpenAI, their newest 4.
• 18:59
  5 model of chat GPT is supposed to be the most emotionally accurate.
• 19:05
  So if you're looking for a, someone that you can converse with and you
• 19:10
  want a more human type response that they're the leader in more of that
• 19:15
  emotional intelligence right now.
• 19:18
  And then, Google, I think is cool.
• 19:20
  I keep wanting to, and I go back and forth between chat GPT and Google's Gemini.
• 19:26
  There's a something that I really I'm hoping that and I think they just they're
• 19:30
  figuring it out right within the last week or so where they're tying it into
• 19:35
  all your past emails, all of your files, all of your search history and everything.
• 19:39
  And learning about you and becoming more of that personal assistant.
• 19:44
  And that's where I feel like Gemini is going.
• 19:48
  They're not there yet, but they're starting to pull all that together.
• 19:51
  Google knows so much about us.
• 19:53
  You know, and we live our lives on our computers and our phones now.
• 19:57
  And so much of that is through Google search and all that.
• 20:01
  I think, when I think about how, what I want AI to do for me, generally,
• 20:05
  would be to be an assistant.
• 20:07
  That person who I can go to and ask them to do things, or ask them to remember
• 20:12
  something about me, or make suggestions based on what it knows about me.
• 20:16
  So that's, I feel like Google's poised to really do that part of it.
• 20:21
  That all makes sense to me.
• 20:22
  Those are most of my daily drivers.
• 20:24
  I don't do too much on Gemini.
• 20:26
  Because I think Google does already know way too much about me and
• 20:28
  I'm a little paranoid like that.
• 20:30
  That said, I'm on Perplexity and I use all of Perplexity all the time and Perplexity
• 20:35
  just rolled out a couple weeks ago.
• 20:36
  They just rolled out.
• 20:38
  The ability to have access to your file stores through SharePoint, which I
• 20:41
  found a lot of value in but that's kind of , my personal, my rundown is I use
• 20:46
  chat GPT for most of my daily driving.
• 20:48
  I use the Perplexity I don't remember if it's Perplexity Teams or Pro but
• 20:53
  I have that tied into my SharePoint data stores so I can ask, pertinent
• 20:57
  questions access information there.
• 20:59
  I do use Copilot actually quite a lot just because I can ask, about
• 21:03
  my schedule, about things that are coming up, I can find documents in
• 21:06
  SharePoint for a lot of our audience, I think that they're almost all of
• 21:10
  their document storage lives in, a file line or neos or cloud Lex or something
• 21:16
  like that which I think it opens up an entire different conversation, right?
• 21:21
  If you're using a case management or a CRM or something like that,
• 21:25
  then you have different needs.
• 21:26
  And I think those need to be, those need to be looked at individually.
• 21:30
  And then I use Claude actually for a lot of writing, if I need to write
• 21:33
  something, I generally go to Claude.
• 21:34
  I do use Google for I use their, what is it Google LM, their Oh, and
• 21:40
  Notebook LM.
• 21:41
  Notebook LM, yep, I use that when I want to take in a lot of information
• 21:44
  learn quickly about a topic.
• 21:46
  I wonder what I don't really mean this as a question, but you're walking
• 21:49
  away in, but the Dunning Kruger effect that AI is going to usher in,
• 21:54
  I think, is going to be profound.
• 21:57
  I've started to see that already, where someone plugs into AI and then
• 22:02
  Minutes later thinks that they're an expert in a field and it's oh man you
• 22:07
  learned this much And it looks so easy.
• 22:11
  And then there's that the rest of the curve right where it's like Well,
• 22:15
  you it's you were able to pick up the guitar and learn how to play,
• 22:18
  you know smoke on the water You got a ways to go Yeah I'm curious what
• 22:25
  that's going to look like within our society as the years go on, because
• 22:28
  , it's such a low bar to get on board.
• 22:31
  I think it just augments what's already happening.
• 22:34
  You see it on social media where people are going off half cocked
• 22:37
  with their opinion that's probably based on very little fact and AI just
• 22:44
  bolsters that it just reinforces it because now they got something they
• 22:47
  can point back to and say, well, look, it's chat GPT says it right here.
• 22:51
  Well, it's still wrong.
• 22:52
  It's not, that doesn't mean that it's fact just because an AI model said something,
• 22:58
  you know, I'm sure your listeners are familiar with the term hallucination.
• 23:01
  It's a real thing and you need to research your answers
• 23:04
  before you, stand behind them.
• 23:07
  Yeah, most of our clients live through the 60s, so they've got a
• 23:09
  different idea of what that means.
• 23:11
  But, uh, they're coming around.
• 23:14
  I appreciate it, Pete.
• 23:17
  Do you want to leave us with what is one thing that a lawyer should do today to dip
• 23:24
  their toe in the space, in your opinion?
• 23:27
  Yeah, if I look back in probably the biggest mistake that I've made
• 23:34
  in implementing AI is Thinking that just because I was the one that
• 23:39
  was learning about it meant that it was going to get implemented.
• 23:43
  And if I had to do it over again and how I'm going to change it moving forward
• 23:47
  is I need to train everybody on it.
• 23:49
  I, we need to whether it's some kind of a platform where you earn a certificate
• 23:53
  or we do lunch and learns or something.
• 23:56
  You have to train the organization.
• 23:58
  Your adoption rate is going to improve dramatically.
• 24:02
  I've worked with a bunch of technical people.
• 24:05
  You would think it'd be the easiest group of people to get
• 24:08
  on board with AI, and it's not.
• 24:10
  And I scratch my head a lot, and just, and actually just getting, preparing
• 24:16
  for this, it kind of dawned on me as I was looking at different things, like
• 24:20
  So I was reading success stories about implementing AI, and a lot of it came
• 24:25
  from just general and broad education.
• 24:28
  And I think that's, that's the key, I think.
• 24:31
  Yeah, I agree.
• 24:32
  It had always been that I would tell everybody, regardless of the position they
• 24:36
  started with us, and they need to learn to be good writers, good communicators.
• 24:41
  And if you can write a good convincing email, then that is
• 24:46
  a skill that you're never going to be sorry that you developed.
• 24:49
  Yeah.
• 24:50
  And I think that to your point, we're at a spot now where if you
• 24:54
  could write a good prompt, that is a skill that you're gonna need for
• 24:57
  the rest of your career, you know?
• 24:59
  Exactly.
• 24:59
  Um, so it'll be interesting to see how that comes along.
• 25:02
  And then uh, I appreciate your time, man.
• 25:04
  It's funny, like right before we got on this, I just got this in the mail here.
• 25:07
  This is a pendant.
• 25:09
  Because I too have been bitten by the bug and like, if it came out, I've tempered
• 25:14
  myself a little bit, but I got the glasses with, you know, I got the glasses with
• 25:18
  the microphone in it and all that stuff.
• 25:20
  I just got this one today.
• 25:21
  So this came out, we'll see how the pendant does, but I appreciate your
• 25:24
  perspective, Pete, your insights.
• 25:26
  This is really helpful.
• 25:27
  I think that the folks listening to this will get some benefit from it too.
• 25:30
  So thank you so much.
• 25:32
  Yeah, my pleasure.
• 25:33
  I love talking about it.
• 25:34
  Like I said, so you're gonna have to cut me off or I'll, or I'll keep going.
• 25:38
  A big thank you to Pete for joining us today.
• 25:40
  You can find Pete on LinkedIn or check out a West Michigan AI Lab Meetup, which
• 25:44
  you can learn more about on Meetup.Com.
• 25:48
  Legal Marketing Radio is a production of LaFleur Marketing.
• 25:50
  Find us at LaFleur.Marketing and on LinkedIn, Facebook, and Instagram.