Anthony DeSimone | Making AI Work for Your Law Firm Without Compromising Security or Excellence

Elise Holtzman: Hi, everyone. It's Elise Holtzman here, a former practicing lawyer and your host at The Lawyer's Edge podcast. Welcome back for another episode.

Let me ask you something. Have you ever typed a question into ChatGPT and been amazed or maybe a little freaked out by how good the answer was? You're not alone.

AI has moved from being something we used to joke about in science fiction movies to something we're using—or at least thinking about using—every day in our professional lives, whether it's drafting documents, brainstorming marketing ideas, analyzing data, whatever it may be, AI is starting to reshape how law firms work.

That's why I'm so excited about today's episode. We're going to dive into how law firms can strengthen and streamline their operations using artificial intelligence. My guest is someone who's truly on the front lines of this transformation.

Before we dive in, today’s episode is brought to you by the coaching team at The Lawyer’s Edge, a training and coaching firm that’s been focused exclusively on lawyers and law firms since 2008. Each member of The Lawyer’s Edge coaching team is a trained, certified, and experienced professional coach and either a former practicing attorney or a former law firm marketing and business development professional.

Whatever your professional objectives, our coaches can help you achieve your goals more quickly, more easily, and with significantly less stress. To get connected with your coach, just email the team at hello@thelawyersedge.com.

My guest today is Anthony DeSimone, the president of You're the Expert Now and a seasoned CPA and CMA, which is Certified Managerial Accountant, with more than 23 years of experience. In addition to offering an array of services that uphold expert financial guidance, Tony is a leader in helping small and mid-sized law firms and other businesses properly use ChatGPT and other AI technologies to boost efficiency, safeguard sensitive data, and enhance profitability.

Tony started at Deloitte, moved through significant roles such as CFO, COO, president, and trusted advisor, and is now a business owner whose passion is turning small businesses into efficient, streamlined operations.

In addition to running his business and being involved at the University of New York at Buffalo as an adjunct professor and executive director, Tony's the co-creator, along with John Remsen, of the Managing Partner Forum’s AI community, a program of forward-thinking law firm leaders dedicated to navigating the AI landscape together.

Tony, welcome to The Lawyer’s Edge.

Anthony DeSimone: Thanks, Elise. It's a pleasure to be here.

Elise Holtzman: I'm really happy to have you here. You and I have worked side by side—I don't know if we'll call it together—but we work side by side as faculty for the Managing Partner Forum, which is where we met.

You have quickly become Mr. Popularity there because of your expertise in AI. As you know, we've done a lot of live polling when we have those managing partners in the room. I think there's a lot of excitement, but also some fear and concern about AI.

So you've worked with so many small and mid-sized law firms. What are you seeing in terms of how they're approaching AI right now? Are they excited? Are they cautious? Are they confused? Are they diving right in? What are you seeing as you talk to law firm leaders?

Anthony DeSimone: They're all over the place right now. There are some law firms that are totally adopting it and jumping all in. Then there's others that really haven't gotten started yet.

Really, the risk with that, it's so important that everybody gets started, because a survey came out just recently, a few weeks ago. One of the survey results was that 35% of employees are now bringing their generative AI tools into the organization. In most cases, the employer doesn't even know it.

So now we're in this messy middle where generative AI is being integrated into a lot of the software that we're already using, and people are not properly trained on how to use it. In law firms, this is really critical because you have now lawyers—individuals in your organization—who are taking confidential information and unknowingly putting it into their generative AI tool.

We also have AI hallucinations that are happening everywhere. You know, Elise, about four months after ChatGPT was released, which was about two and a half years ago, there was this well-known story—all the lawyers heard about it—where a lawyer used ChatGPT to cite fake court cases. ChatGPT just created these court cases out of thin air and got into a lot of trouble for it.

Well, this is happening every week now because the lawyers who are using it don’t understand how to use it properly and don’t understand the risks of things like AI hallucinations.

Elise Holtzman: What’s remarkable to me is that after that poor guy did what he did—it was in all of the legal media and in some other media as well—that other people continued to do it. When I first saw it, I thought, "Oh, thank God. Poor guy got caught, but now no other lawyer would ever do that." You're telling me it's happening all the time?

Anthony DeSimone: All the time.

Elise Holtzman: So people aren't understanding how to use it and aren't paying attention. So why are people waiting? What are some of the misconceptions or legitimate fears, perhaps, that lawyers or law firm leaders have about AI?

Anthony DeSimone: They just don't know how to use it. It's the fear of getting started sometimes and some of the stories that you've heard.

So a lot of times people are just not getting started because they don't really know where to begin and they're afraid to get started based on some of the stories that they heard. But that's the biggest mistake anybody can make, because every presentation I give, I make sure they hear that ChatGPT won't replace you. The people who use it will. The law firms that use it will replace the law firms that don't.

Nobody goes to work and says, "Hey, should we turn on the internet today and use it?" That is how people are going to be thinking about generative AI. It's going to be so obvious. It's going to be surrounding us just like the internet. It's going to be everywhere that you won't even think twice about it. But getting started sometimes is the hardest part.

Elise Holtzman: You mentioned a few minutes ago that some of the tools that law firms are using already incorporate generative AI into their platforms. It sounds like you're suggesting that many law firms are already using AI, they just don't even realize that that's what's happening. Is that a fair assessment?

Anthony DeSimone: Well, yes, it's possible. Google, for instance, has Gemini, and now that's integrated—that's their large language model—it’s integrated into Google products. Microsoft has Co-Pilot. Co-Pilot is another competitor of ChatGPT. That’s integrated in the Microsoft products.

Some of the law software like Lexis, Westlaw, things like that, they have an AI component now to it. So now you have this option to purchase these AI to go along with whatever it is, the software that you have.

Some of them don't even realize, but it's just been automatically integrated into it and they're using it. Now, the difference between somebody using it well and somebody using it incorrectly is being trained and really truly knowing how to use this technology.

Elise Holtzman: It occurs to me that, listen, I started practicing a long time ago, and I remember a time where I was at a big law firm, and they suddenly announced that you were no longer allowed to send paper memos around to land in the inbox that was sitting on your desk. People went out of their minds.

I mean, “What do you mean we have to use email? What do you mean we're not going to be able to send paper memos around?” The idea of sending paper memos around is so utterly ludicrous at this point. We just had to accept the fact—or, I mean, I was very young, so it was easy for me—but people just had to accept the fact, as you said, that email was here, the internet was here, this was what we were going to be doing. And the world changes. You can't get left behind.

Can you share some examples of how law firms are currently using tools like ChatGPT or other AI technologies in practical ways in their law firms? What are law firms doing with this stuff?

Anthony DeSimone: Well, first of all, they're doing some of the basics where they're just having it edit for grammar. I mean, really, everybody should be doing that. Right out of the gate, that's the first use case. I've never communicated better. I've never sounded smarter. That is definitely what everybody should be using it for.

Lawyers are using it for assistance in writing, where now there are features like on ChatGPT where you essentially have an AI assistant working directly with you whenever you're creating a document, side by side, essentially just speeding up the process.

There are what's called deep research functions now in these large language models where it can allow a lawyer to do a much deeper research on any topic, where it'll cite the work, it'll source the work, so that they can check it and make sure that there are no AI hallucinations or bias to go along with it. Those are three very quick and easy ways that really all lawyers should be using it.

Now, they take it to a next level when they start using it, and that's the beauty of this technology, by the way. I could spend hours talking about use cases, but the use cases that you're going to discover that are best for you are going to be discovered by you.

The only way you get there is by understanding and using this technology so you can start connecting the dots like, “You know, this could do this, and now it can do this for me.” And it's more of a workflow type of process where it goes from one to the next to the next.

We have lawyers out there—there are generative AI tools, one's called Fyxer, where they're using it to answer their email. So think about it. It's got this whole workflow situation where it'll read your email, it will determine which emails need to be responded to, it will create the drafts of those emails, and all you have to do is go into your drafts section of your email and review the drafts it created, make the adjustments that you need to make, and send them off.

We're starting to see more and more of that type of workflow.

Elise Holtzman: That's fascinating to me, partly because one of the things that I do with my executive assistant is that she will—I just call it teeing things up. So she will sometimes tee up email responses for me, leave them in my drafts folder, and then I go back and I refer to it as zhuzhing them up.

I'll make sure that it's what I want to say, I'll zhuzh it up. I'll put in some personal comments for somebody that I know well, that kind of thing. It's incredibly helpful.

The idea now that we can have AI do something like that, I mean, think about the time saver that that can be. Even just the mental clutter of feeling like, “Oh my God, there are all these emails I need to respond to,” that's a huge challenge for most lawyers, is they become overwhelmed by the rush of emails that they're constantly finding in their email box.

Anthony DeSimone: Yeah, and I've been using it now for a few months and it saves at least anywhere between two and five hours a week. There's no doubt, because I don't know about you, but sometimes I'm way overthinking some of these emails and then trying to write it properly and everything.

Sometimes I get a draft that's almost perfect. It takes two seconds to just clean up and it's done. That's so helpful.

Elise Holtzman: I've been shocked too because one of my skills—I don't have that many of them—but one of my skills is writing. I've taken—just to see how it works—I've taken an article I've written or something like that, a blog post, and I've put it into ChatGPT or Claude or one of those and said, “Hey, here's an article I'm writing. Can you make it better without getting rid of any of the important content?” You know what? Every single time, I've never been disappointed. Every single time, it's done a better job.

So Tony, lawyers are understandably concerned about confidentiality and data security. I think this is one of the things—and correct me if I'm wrong—this is one of the things that lawyers complain about or they're afraid of. Because I had asked you about what their fears are. One of the things you said is they just don't know how to use it. They don't know what it's capable of. They don't know how to use it.

I think there's also this fear of the confidentiality piece. Because if nothing is drummed into our heads in law school and when we get to a law firm, it is the confidentiality piece. So how can law firms use AI responsibly without risking sensitive client information?

Anthony DeSimone: Yeah, when it comes to lawyers, what they're afraid of is the security issue as well as the AI hallucinations. They don't want to be the next person putting fake court cases out there.

What needs to be done there—first of all, when it comes to security—the first thing that I always say, the piece of advice, because you want to be as conservative as possible in using these tools is: don't put anything confidential or proprietary into any of these generative AI tools unless you're sure that they have the level of security and encryption that your firm allows. Period.

Now typically the follow-up question is, what is that level? What's the acceptable level? The way I tend to answer that one is that, you know, basically most people are using Microsoft 365, let's say. So you never think twice about putting confidential or proprietary information in that Word document or that Excel spreadsheet. You don't even think about it. You just do it.

So at the least, these generative AI tools should have that level of security and encryption. So if you don't know if they do, then you probably shouldn't be putting confidential or proprietary information. I mean, that's the rule of thumb.

So what I suggest that law firms do—organizations in general—is that you have to define what is that level of security and encryption that we need to have in order to allow that confidentiality to be put in there, confidential information, things like that.

Then once you've got that lined up, now you can share with the team: these are the generative AI tools that are approved, and these are the ones you can and cannot put anything confidential in.

Then with AI hallucinations, I think this is the biggest failure of people who are using this tool because they're not properly trained. But generative programming, these generative AI tools were created with generative programming. And basically, the way generative programming works is simply this: if there is a lack of information, the generative programming will kick in and fill that gap of information with what it believes should be there.

Sometimes it gets it really wrong. I mean, it's a total hallucination. That's what they call it—an AI hallucination—when it gives you false information, but it appears that it's accurate.

People don't realize that because of generative programming, it is interwoven into the DNA of generative AI. You're never going to have generative AI without the risk of AI hallucinations. It's gotten better over time, but it's never going to go away.

This is why it's incredibly important that anything you create on generative AI tools needs to be checked, needs to be edited by a human before it gets sent out.

Elise Holtzman: I think that's the thing that people need to remember, we're not replacing human beings here. We're making their lives easier. We're making things more efficient.

We're able to serve the client more effectively when we are focused on the real issues that matter and not focused on the things that can be done easily, efficiently, and less expensively.

What do you think the law firm leader role is here? How can law firm leaders encourage experimentation and innovation without creating chaos, first of all, and then also without maybe creating resistance or somehow overcoming the resistance among attorneys and staff?

Anthony DeSimone: Well, there are over ten states now that have offered opinions or guidelines, including Washington, D.C., and the ABA. They're all saying something very similar when it comes to lawyers and the use of generative AI.

First of all, lawyers need to understand the risks and capabilities of the generative AI tools that they're using. As a supervisory lawyer, a supervisory lawyer needs to understand what their team is using and has to understand, again, the risks and capabilities of those tools if they're to supervise the people that are using them. So there's an obligation there.

Really, the first step is proper training, getting into the habit of using these generative AI tools. The good news is this, when it comes to learning this technology, this knowledge, this generative AI knowledge, translates over very well from one tool to another.

So if you master ChatGPT—which is probably currently, in my opinion, still the best large language model out there, multimodal, you can create images, you can do all these other things with it—if you master that, then you're going to kind of "understand" how these large language models think. And it's easier for you to go and jump to other tools and use those because you'll kind of get them even if you haven't ever used them before.

Elise Holtzman: That makes a lot of sense to me because I first started with using—and I'm not a law firm. I mean, I practice law, but I'm not a law firm. Obviously, I'm coaching and training and consulting firm. So the confidentiality piece is a little different for me. That part's easier because I can easily just put things in and never use anybody's name or law firm or anything like that.

But I had to play with it. And again, going to date myself, but I was like the last college class that went to college with a typewriter. Okay, so, you know, word processing was just a thing. And two years later, my husband showed up in college—of course, he already had that stuff because that was exactly when it was all happening.

So I don't consider myself an intuitive user of technology. I consider myself somebody that has to learn it. But it doesn't come naturally for me.

When I went from ChatGPT to use another one—I wanted to try Claude—you're right, I just got in there and right away I understood how to use it. The other thing is that I like what you say about experimentation. We have to be willing to try things, and it doesn't necessarily have to be on a bet-the-company litigation.

It doesn't have to be something really important, but getting in there and trying things has made me a lot more confident with the tools. And also has resulted in a couple of funny situations where I actually got into an argument at one point with ChatGPT and got so annoyed with the whole thing that I actually basically said, "Stop gaslighting me."

Because it was a ridiculous situation where I was asking it to do something like putting people into groups and making sure that Tony and Elise weren't in the same group, and Sally and Jimmy weren't in the same group. And it kept saying, "Hey, I've done all these groups for you," and it kept putting it in front of me and it kept being wrong.

I finally said to it, "If you can't do it, just tell me you can't do it and stop gaslighting me." People are laughing at me now because I was in an argument with ChatGPT, but it was really helpful for me to see what it really didn't do particularly well, which was sort of math-like.

Somebody said to me, "Elise, this is on a large language model. You know it doesn't do math and sets particularly well."

So I do think that that's important to realize. Yeah, you do have to play with it. You do have to understand how it works, and you can't expect it on the first day either to ruin everything or to fix everything. It's not one or the other. There's a lot of gray area in between.

Anthony DeSimone: Yeah, and you need to understand what their capabilities are, because these are large language models. They were trained on text. They were trained on words. They weren't trained on numbers.

So they are fantastic at helping you with your grammar and just documents and all that great stuff. They have a PhD level in that. But they have, like, a fourth-grade level in adding and subtracting. I mean, they're terrible with that.

From my perspective as a CPA, you don't know how badly I wanted to be able to create these amazing Excel spreadsheets and just do all this. I've been testing this because I'm hoping one day I'll be able to do it. But just a simple request of adding 30 rows of numbers, forget about it. I mean, it's not even—you can't trust it because it wasn't trained on numbers. It was trained on text. It's not supposed to be great at numbers.

Just like for image makers, if you're familiar with some of the image makers out there, for the longest time, I mean, they were trained on images, not text, so for the longest time, when you would ask an image maker to create something with some words on it, it couldn't spell. It just made a mess of the words and it was like, don't even bother creating images with words in them.

They've gotten better, to the point where they're almost, you know, able to spell all the time perfectly. But once again, another example where image makers were trained on images, not text. So they're going to have a problem with that.

Elise Holtzman: So given all that, where do you see all of this headed in the next three to five years?

I guess improvements obviously are going to happen, and presumably they're going to happen pretty quickly. You talked about some of the things that law firms could do with AI. What are some of the things that we haven't discussed that law firms already are doing with AI? Where do you see this going? What do you think they're going to be able to do with it in the next few years?

Anthony DeSimone: First, back to something else that you said, how it's not going to replace people. I agree with that. I think it's not going to replace, it's going to enhance the people that are willing to learn this technology.

The ones that are not willing to learn this technology, it's almost like somebody trying to compete with a typewriter with today's technology. It's just not going to happen.

So where this is headed is that we are going to see, very soon, humans and agentic AI—basically AI agents—working side by side in a law firm doing similar tasks. And then eventually—and when I say eventually, I'm talking 18 months maybe—you'll have human lawyers in charge of several agentic AI.

Then somewhere down the line, maybe another year after that—so we're only talking two and a half maybe years—you’ll have agentic AI, AI agents, that will be in charge of humans. Human lawyers.

Elise Holtzman: I mean, that's where people get freaked out, right? That's where people really get freaked out. Should we be freaked out by that? I mean, I feel like we should, but maybe I'm missing something.

Anthony DeSimone: I don't think so. Well, the genie's out of the bottle at least, so I don't think we have a choice but to continue on this path.

I think for those who truly learn how to use this technology, you're going to be very excited about where this is headed. Those who don't, I mean, they're the ones that are going to be in trouble. But I know I'm very excited about it. I would love to have a full organization of AI agents. I'm working on that.

You know, Sam Altman, who's the CEO of OpenAI, which is the creator of ChatGPT, he said, and he said this about, I don't know, eight months ago now, that within 18 months, there's going to be a ten-person, one billion dollar valuation company.

He said he's making bets with his tech friends as to when there's going to be a one-person, one billion dollar valuation company. That's because of what he sees coming with all these agentic AI that could do so much for us.

Elise Holtzman: Now, when you say, you said before that you do foresee that there will be an attorney running ten AI agents. What does that mean? Are these—these are different AI agents that do different things? So you have an AI agent that does one thing and an AI agent that does another thing?

Because to me, when I hear AI, I think of it as a thing. You're talking about it as being divided into sections. The only thing I can come up with is that that means there are different AI agents doing different things. Am I right about that?

Anthony DeSimone: What I see coming are AI agents that can do everything. Similar to, like, you know, a new hire into your law firm. And you'll have AI agents that will be doing the same thing, where they're going to jump from one app, one piece of software in your firm, to another, to another.

And, you know, review the email, determine which documents need to be created, now go to Word. It's going to jump over to Microsoft Word, create the document. Then it's going to jump over to another—Lexis or Westlaw—and do the research, or whatever it needs to do, until you have the final product. I mean, that's agentic. That's where it'll essentially be a robot that's doing the same thing as humans.

Elise Holtzman: Right. So maybe you have an agent working on a couple of different matters, or a matter in the same way that you might have an associate or a junior partner working on a number of different matters?

Anthony DeSimone: That's it.

Elise Holtzman: Let's talk about young people for a minute. I know that you're an adjunct professor at Buffalo, and my daughter is a summer associate right now. She just started a summer associate position at a very large firm.

The first week of training included training on AI. The assignments that she was given, she’s being instructed to use AI and how to use AI. What are you seeing among young people? Are you seeing that they're just adopting this the same way that young people have always adopted technology?

It's just like part of the wallpaper, they get used to it and they immediately adopt it. How do you see them being different as lawyers than what we're used to?

Anthony DeSimone: Well, there was a recent survey that came out that said that millennials and Gen Z, many of them are actually retaliating against generative AI tools. Part of the reason is that they see these tools as a replacement for them.

And again, I feel that that's a lack of training once again. Because if they understood how these are going to enhance your ability, then you would just embrace that. You would learn it and embrace it.

But really, for the younger generations, there is some concern. Then there's others that just jump right on it. They are just far better with technology, and it makes a lot more sense to them.

So it's kind of a combination of both, right now, Elise. And it's great to hear that your daughter's firm is teaching right out of the gate because, man, they are so far ahead.

I mean, we are still early. I'm speaking on generative AI weekly and there's still more people learning how to spell ChatGPT than know how to use it at this point. So we are still early. So it's great to hear that she’s doing that, especially in a larger firm.

Elise Holtzman: Well, and you and I have had the experience of doing live polling in a room of more than a hundred managing partners. I guess you put this together with John Remsen. You asked the folks in the room who's using and who's not using it.

The numbers were very, very low. These were small to mid-sized law firms, anywhere from five lawyers in the firm to up to 180 lawyers in the firm. The adoption was remarkably low.

Anthony DeSimone: Yeah.

Elise Holtzman: That wasn't surprising to you?

Anthony DeSimone: No, that was normal. Unfortunately, what's more concerning—and this was unfortunately normal too—is the very low percentage of law firms that had an AI usage policy.

That's getting really scary and risky. Because, like I said, so many people are now bringing their generative AI tools into the law firm, into the organizations, even when the organizations have maybe a no-AI rule or whatever, they're doing it.

And if they're not trained properly, and if you don't get ahead of this, there's going to be some form of a breach.

Elise Holtzman: I think people were shocked to hear that. I know I was shocked to hear that, although thinking about it, I shouldn't have been shocked.

I mean, everybody's walking around with a computer in their pocket. Everybody brings their cell phones to work. They throw the phone on the desk. You know, the younger people are barely using it as a phone anymore. They're really using it as their connection to the world.

So I don't think we should be surprised that people are using AI tools because they're starting to see the value in them. So if the firm is behind, you're right, they're going to get themselves in big trouble, partly because also this is becoming an ethical issue.

Well, it's already an ethical issue, whether the states say it is or not, and states are already diving in and saying, "Hey, this is an ethical issue. You have to understand technology enough to be able to protect your client's data and use it responsibly."

We saw that with other forms of technology over the years, where the state bars had to get involved right away and come up with policies for this stuff. So what should a law firm be doing? Even if they haven't yet started using AI tools in a way that is really becoming part of the fabric of the firm, what are the first couple of things you think law firms should be doing immediately?

Anthony DeSimone: Well, first of all, create an AI usage policy. The next thing that I suggest is create a beta team—an AI beta team, is what I call it—where you have at least one person from every department coming together, meeting every other week, and also connected on your Teams or something so you can share ideas in between the meetings, but focusing on generative AI tools.

We're testing. We're just trying to figure out how we can use these in our law firm, the best ways we can use it in every department. As we're doing this, we're building out a generative AI usage policy so that we can determine: here are all the safe ways to use it. Here are all the safe tools. Here are the tools that you don't use. Here are the things that you never put into any generative AI tool.

Just build that out so that you have now a team of people that are getting better and better at using it so that then you can start to phase the training to your whole law firm. Because if you take those steps, you're going to be able to speed up that process and get everybody on board, trained properly so that you can make sure that your generative AI is being used safely and you're not creating things like AI hallucinations and bias.

Elise Holtzman: A couple of things are coming out for me. One is you mentioned that somebody from every practice group department in your firm should be involved. So I want to just be clear with the lawyers and the law firm leaders who are listening to this, we're not just talking about practice groups.

We're talking about accounting and operations and marketing and professional development and all of the different departments. It's not just for the lawyers, it's for everybody.

Anthony DeSimone: This is going to be a game changer and create efficiencies in every industry, in every department. And to avoid those groups is a big mistake, especially for this type of technology.

Because some of the back-office type roles, departments, this is where this technology is going to create some of the greatest efficiencies.

Elise Holtzman: Can you just give me an example of that? If you see in a marketing department or the people who are running operations or the accounting function, what are some of the things—just, if you have anecdotes or just ideas—of what are some of the ways in which this is going to improve efficiency for those areas?

Anthony DeSimone: Let's do something really simple that everyone struggles with, finding qualified people of every level at any title, right? You can use ChatGPT.

ChatGPT has a feature called Deep Research, and I've tried this and I know it works, where you just prompt ChatGPT. You upload the job description: "Here's the job description for our law firm. We are looking to hire this title." And I want you, ChatGPT—so we're now prompting ChatGPT—scrape the internet, scrape LinkedIn and other social media posts, and find ten qualified people that are either looking for a job or maybe they're not even looking for a job, but they fit the criteria.

And give me that information in table format with their email address or their link to their LinkedIn so that I can reach out to them.

Elise Holtzman: Yeah. The headhunters are going to love that.

Anthony DeSimone: Oh, totally. You know what? For anybody who's—everybody has to hire eventually—that should be the first step that everybody takes from now on because you'll find that you're going to get very qualified people.

You know, I tried this out. I have a client who's looking for a controller, and it's a plumbing company in this case. They were looking for a controller and I did this prompt. And the first person that came up—because I asked it, I said, "Look for people who are looking, but also look for people who fit the criteria that aren't looking"—and the first person that came up was a person that I helped hire in a competitor plumbing company about three years ago.

And I know she's fantastic, and I'm like, "This is great. This is a good example of how that really works."

Elise Holtzman: Yeah, that's a great example. You also mentioned putting into place an AI usage policy. We know those things are important, and more important than that is educating people about the policies.

If you're a law firm leader and you don't even know where to start, where do you get the advice you need in order to create a policy like that? I mean, obviously from someone—obviously, from someone like you, Tony. I presume that this is something that you're advising law firms on all the time. Where can they get started?

Anthony DeSimone: Well, I created a chatbot—surprise, surprise—that builds these first draft AI usage policies. So I could certainly share the link with you, Elise. You could share it with anybody who's listening here. But I highly recommend you start there.

What I did with this chatbot, just so you have an idea, is I uploaded all the opinions and guidelines from all of the states, including the ABA. Plus, I've uploaded example usage policies, so it knows what structure to kind of create. And it uses all that information when it builds that first draft AI usage policy.

I highly recommend you start there. That's why it's so important to have this beta team and having this group that truly begins to learn how to use it, because that's the only way you're going to get that final AI usage policy. When people start getting it and then they start understanding, "Oh, we should never use it for that," or "These tools are safe," or whatever.

Elise Holtzman: I love that we're using AI to generate the AI usage policy. I mean, that's the perfect example.

Anthony DeSimone: You know, that's another thing that I want to make clear to everybody, that 100% of everything I create has an AI complement to it. You know, and that's what everybody should be getting to, where you have that muscle memory, where you're not thinking about it anymore. You just know that this tool is going to help me speed it up. You're just going there. That's where everybody needs to eventually be.

Elise Holtzman: Just one quick example from my own experience is that I'm kind of getting to that point now. Again, this is not something that came naturally to me, but I decided to dive into it.

What I do now is I have a lot of material. As more senior lawyers do, you've done a lot of different things. You have a lot of experience. One of the things I do is I've been feeding—if I'm doing a program—I feed my content into it.

I say, "Look, this is my philosophy on this. This is a PowerPoint that I've done. This is an article that I've written. You know, these are the things that I think are important." It's done some things for me, like for the Managing Partner Forum conference that we just did down in Atlanta, I asked it to help me create handout materials because we were putting them into a bound book that everybody was going to walk away with. I was very, very deeply involved in figuring out what was appropriate for the group. But it really helped me take what I already have and put it into a format and ask some questions and prompts for people that could be useful to them.

So I am getting into that habit. I think that's great advice for people, Tony, to realize that this is going to become part of their everyday existence. It's not going to be just using it occasionally. It's going to literally be a tool that's in everyday use, the same way we're using our desktops and laptops and phones and Google and all of that sort of thing.

Anthony DeSimone: Right on. That's exactly how it's going to be used.

Elise Holtzman: Tony, as we wrap up our time here together today, I want to ask you a question that I ask all of my guests at the end of the show.

There's a phenomenon called the curse of knowledge, where experts sometimes forget that what is so obvious and natural to them is not at all obvious to other people. When it comes to integrating AI into the daily operations of your law firm, particularly now, at this time in its evolution, what's a principle or piece of advice that may seem obvious to you but is important for people to hear?

Anthony DeSimone: Have it open all the time. I think that that's probably the best piece of advice I can give for beginners, is have it open on your desktop. Have ChatGPT, or whatever your generative AI tool of choice, open on your desktop all the time.

Similar to how you have Google open, the search engine is always there. Do the same with ChatGPT. When I have my computer turn on, it turns on to ChatGPT. So just having it open all the time helps create that muscle memory.

It's front of mind, and it just gets you into the habit of that process of thinking, "Can I do this faster? Can I do this easier? Can I do this better with ChatGPT?" Before you know it, you'll get to the point where you'll truly understand the power of this technology.

Elise Holtzman: Fantastic. All right. Well, thank you so much for being here, Tony. It's been really interesting, as always. And thank you to our listeners for tuning in.

If you've enjoyed today's show, please subscribe, rate, and review us at Apple Podcasts, Spotify, or your favorite podcast app. In the meantime, be bold, take action, and make things happen. We'll see you next time.