Dennis Kennedy is an award-winning leader in applying the Internet and technology to law practice. A published...
Tom Mighell has been at the front lines of technology development since joining Cowles & Thompson, P.C....
Published: | April 4, 2025 |
Podcast: | Kennedy-Mighell Report |
Category: | Legal Technology |
Not a week goes by without another new AI tool hitting the market, and this flood of product choices can be incredibly overwhelming for any consumer, lawyers included. So, how do attorneys figure out what they need and which tools are trustworthy? Dennis and Tom talk through the many categories of AI tools and explain how learning to use major LLM tools is essential in understanding AI’s uses and capabilities. They discuss ChatGPT, Claude, Gemini, and Copilot to give legal professionals a sense of how to engage effectively with AI tools and then offer their take on how to narrow down the multitude of AI choices for your law firm.
Later, can Tom guess Dennis’s favorite and most-used AI tool of late? Tom recruits ChatGPT to help him guess!
As always, stay tuned for the parting shots, that one tip, website, or observation that you can use the second the podcast ends.
Have a technology question for Dennis and Tom? Call their Tech Question Hotline at 720-441-6820 for the answers to your most burning tech questions.
Links Mentioned:
LTH GenAI Legal Tech Map: February 2025 | Legaltech Hub
Outlook Calendar in Microsoft Teams
Personal Strategy Compass – by Dennis Kennedy
Special thanks to our sponsors GreenFiling and Verbit AI.
Announcer:
Got the world turning as fast as it can hear how technology can help legally speaking with two of the top legal technology experts, authors and lawyers, Dennis Kennedy and Tom Mighell. Welcome to the Kennedy Mighell report here on the Legal Talk Network
Dennis Kennedy:
And welcome to episode 388 of the Kennedy Mighell Report. I’m Dennis Kennedy in Ann Arbor,
Tom Mighell:
And I’m Tom Mighell in Dallas
Dennis Kennedy:
In our last episode, we spoke with Michael Kennedy and Amy Brookbanks of Shaw Goddard as part of our Fresh Voices on Legal Tech series. Highly recommended if you haven’t listened to it already, lots of great information and insights. In this episode, we want to talk about AI again, but in the context of how you might compare and choose what AI tool or probably tools that you want to use. Tom has really, really wanted to do this topic and with less of new AI tools appearing lately, we thought it might be good to focus on this topic. Tom, what’s all on our agenda for this episode?
Tom Mighell:
Well, Dennis, in this edition of the Kennedy Mighell report, we will indeed be talking about selecting or the right way to go about selecting an AI tool or tools. In our second segment, we will answer a question from chat GPT asking as a stand in for you, our audience. And as usual, we’ll finish up with our parting shots, that one tip website or observation that you can start to use the second that this podcast is over. But first up, choosing an AI tool that works best for you. Like you said, I have been wanting to talk about this for some time, although I really wanted to talk about this a long time ago and I feel like I don’t know that right now is, I think this would’ve been a better topic a while back, but we’re going to soldier through anyway. But I’ve been doing it because I was seeing that not a week was going by without some new AI tool coming out.
And I have to believe that for those who are not regularly using a tool or who haven’t started to do anything or really thinking about it, that it was incredibly confusing just trying to figure out which ones to try, why you’d want to try it now that we’re several months down the road from when I originally wanted to do this. Even more tools or even more functionality is out there, I guess is maybe the better way to put it. So we’re going to try to offer some practical advice on the things to think about. We’re not necessarily going to recommend any tools. We’re really going to talk about the things to think about. So Dennis, what did I miss in that setup?
Dennis Kennedy:
Oh, I think we were really going to recommend some tools. Well, it wouldn’t be fun without doing that. Yeah, I think that a lot is happening these days and this episode as with any episode on AI, is likely to be somewhat obsolete probably by the time it gets released because so much is being released all the time. And we’ll talk a little bit about that. So this is a topic where I think you have to develop frameworks and approaches to think about this. I would say the good news to me is that if you kind of sat back on AI for a little bit and now jump in, now the level of the tools is really much higher than it was even say three or four months ago. So if you feel like, oh, I didn’t get in there and learn it and pick a tool three or four months ago, you’re kind of in a better place to do that. So that’s an interesting almost paradox about where we are now. But I think it’s worth talking at the beginning time about how many tools are being released almost. It used to be weekly now it’s almost daily. So what’s your thought on that as you see this kind of fire hydrant of AI tool releases coming out to us on a regular basis?
Tom Mighell:
Well, I mean I think we have to be able to separate out the types of AI tools that are being released. I mean, there are the law based ones which tackle specific legal tasks, research, drafting of legal briefs, analysis, things like that. There are AI tools that will take notes in your meetings or that will help create images for you. I want to spend probably most of our time today talking about the general purpose, large language models that are really more for general use that don’t have a specific use case in mind to do that. Because I think that for people who are still trying to decide, I think that’s the best place to start. You could certainly dive into a very targeted tool if you want to go and buy a tool that will analyze your depositions and help summarize them for you. That’s all well and good, but I think that gives you a narrow view of what the tools can do.
I think that I would recommend that you start broad to see what they can do and then focus in on the things that are useful for you. That’s really how I would start. I think that telling people as the script says, is to look at the legal tech hub 400 legal AI tools chart. I think that’s pointing a fire hose at them and just saying, go pick one, and that should be easy. And I don’t think that’s the fair way to go about it. I think that by talking about the big three or four or five large language models that are out there, I think we do our audience a better service by talking about that because it makes the choices a little more straightforward at the beginning of the process.
Dennis Kennedy:
So I think that the legal tech hub, legal AI’s tools chart is really an interesting story of the times that we live in. So they did all this great work on it and I guess in disclosure, I write a column for legal tech hub on law department innovation. So I think they did a great job and they put out this chart of 400 legal AI tools that they had identified and that people had identified as being legal AI tools. And they got a lot of pushback right away. And the pushback was, and this reflects our times, people said, you forgot this. You didn’t list my tool, you didn’t do this. And so in a couple weeks later, they’ve released a revised version of it, which is now at 505 legal AI tools and divided into categories. And I would say, Tom, and you kind of alluded to this, if you give me 505 options, you’ve essentially given me zero options.
I freeze up. I don’t even know what to do with that. And so that’s why I look at frameworks and say, how would I start to think about this? And then it’s compounded because every time a new tool comes out, you have a bunch, I dunno if a bunch of ’em, but you have the experts who seem to say within 24 hours that this tool is the best one ever and the best one that you should use. And I don’t even know. Or they get it barely even trying something before people are saying it’s the best one and then a few days later they find something else. That’s the best one. And it is just super, super confusing. So that’s the problem that we face. And there’s part of me that says there’s a famous photographer, chase Jarvis who said the best camera is the one that’s with you.
So it’s like maybe the approach in these kinds of times is just to pick one that makes sense to you and this goes in time. I think this takes us to the general purpose tools. Just pick one and learn to use it. Well, and to me, I will always say, and I’ll emphasize in this podcast that it’s not so much the tool, it’s like how we’re using it, how we prompt those kinds of things. And so in a way that best tool I want is the one that makes it easiest for me to learn to prompt really well.
Tom Mighell:
Okay. Well, it really comes down to four major AI tools that people need to think about. So pretty straightforward. So chat PT, the one that’s been taking all the oxygen out of the room, that’s by OpenAI, Claude, which came along right after that and has been making significant gains since then by Anthropic Gemini, which we’ve talked about on the podcast tool. And then copilot, which you all should know a little bit about. It’s Microsoft’s tool that they’re using within Microsoft 365. So I think those are the four big ones that we want to talk about. And I think, Dennis, what is your sense of what you’re out there of? What are the tools that people seem to be using most often? And I’m not asking you your question because that’s actually what we’re going to talk about later, but what do you think that others, what is your sense of what people tend to be using right now?
Dennis Kennedy:
So what we’re seeing to the extent that we have surveys that seem credible is that chat GPT is significantly in the lead that copilot is often used because in enterprise situations that’s the one that’s available. Then you see some people experimenting with Claude, I would say much lower percentage. And then Google stumbled so badly out of the gate that a lot of people have stayed away from Google Gemini. But it’s the improvements it’s made in the last couple of months. Even the last couple of weeks have really changed my thinking about the Google tools. And then if you consider Google Notebook LM as part of the Google Gemini tool set, that I would say that’s another piece of it that’s attractive, but I would sort of rank it in those orders, Chad, GPT, copilot, Claude, and then Gemini and back in the days that people rated hit records, I would say that Gemini might have the upward arrow beside it is one that’s I likely see to see grow and chit. GPT might be the one who kind of falls back just a little bit, but I still think it’s pretty much the standard.
Tom Mighell:
Okay. Well then let’s, maybe what we do is in that order that you ranked it, not sure that I fully agree with the ranking, but I’ll go it. I’ll go with it for now. Let’s talk in general about each one of them individually about our thoughts about each one. So chatt pt, that’s the one that people have heard the most about. It’s still what I would consider to be the default, the one that people think about the most. Due to early advantage, it was the first out of the gate, it was the one that everybody was talking about. It really captured the swell of the hype cycle, if we could consider it that way. It is easy to use, it’s versatile as we’ll talk about a little bit later. It has a lot of different flavors that you can choose from to do allegedly a lot of different things.
So it has taken a lot of the, like I said, the oxygen from the room. I think it’s well suited for general question and answers. It’s good for a brainstorming content. It has a good, I think conversational interaction. I’ve been able to have good conversations with it. So I think it’s good in that aspect. I think that if we think about the limitations that chat GPT has, I would say that still the risk of hallucinations is still a real risk. The thing that I wish some of these tools were better at were the real time data. Unless you are using plugins or you’re doing web browsing or some other thing with it, you’re going to have outdated information that it’s going against. I think that unless you’re using one of the turbo flavors of it, then you may have a limited context size from what you’re dealing with with chat GPT. So there are some limitations, but I think to be honest, it’s the one that I use more often than I use perplexity as my search engine. But I would say that chat GPT is a tool that I probably use most often, mostly because I don’t want to have four different subscriptions to it, and that’s one of the ones that I’m using. So
Dennis Kennedy:
I’ve been using chat GPT for two and a half years. So it is the standard for me. I think it’s a great one to learn on. I think it’s a great place to learn prompting. It has a number of flavors that allow you to do more sophistication. You have to pay for it. You’re not even talking about AI if you’re not paying for these tools. So I think it’s the good basic one. People are familiar with it, and I think at this point what you’re trying to do is to find a tool that you can learn to prompt really well. So there is a part of me that says if I’m choosing a tool, I probably lean toward picking one that’s like my primary platform and this is what I would recommend for people. So even my students who have the access to Lexus ai, Westlaw AI tend to, when they get disappointed with those tools, tend to gravitate toward a paid version of Chad GPT. So if you kind of said that I had to recommend one place to go to, that’s the place I would say for the bulk of people, put down the 20 bucks a month and just start working with you GPT. So Tom, next up would be Microsoft Copilot. You usually are the Microsoft fan here. So your thoughts on copilot,
Tom Mighell:
A little sneak preview. We may be having an upcoming podcast on Microsoft copilot shortly, but I will tell you it’s not my favorite. I think that it has a place and I think that there are definitely use cases for it, but I would argue that it’s probably not as strong as some of the other tools we’re talking about today. One of the thing that makes it good is it has a very, very strong integration with the Microsoft 365 apps is that you can use it within most if not all of the apps that you would use, word, Excel, PowerPoint. You can use it separately. You’ve got access to a little chat feature with copilot that you can just use it the same way that you use chat GPT or the other tools. So it’s a natural choice for productivity heavy environments. It is a great productivity artificial intelligence engine.
It’s ideal for document creation, it’s good for summarizing meetings. It will sit in all of your teams meetings and summarize them can while you’re in the meeting, you can query copilot and ask questions like, how’s this meeting going? What are some of the issues you see that we need to still resolve? It’s really terrific about doing that. It can do good workflow automations. So there are good things within the Microsoft 365 sphere that really it’s going to be better than anybody else, but it really, its effectiveness depends on how proficient you are with Microsoft and how if you are a Google shop as a lawyer, this is not the tool for you. You’re going to want to think about something else. There are questions around data privacy that are some issues there, but I think it’s a good tool. I don’t use it the same way that I would use chat GPT because I found that the results that I get there just aren’t as satisfying to me as some of the other tools.
Dennis Kennedy:
And so for me, the thing on copilot is I think if you’re in an enterprise setting, this is likely to be the tool that you’re allowed to use and that becomes a big thing. There are still firms and organizations that don’t allow you to use chat GPT even at this point. So copilot becomes the option and then because it is the option, it’s a solid choice. I think that the Microsoft copilot for me is kind of really confusing the way that it’s used and there’s a couple different flavors of it and that it is an over, and I mean this in the most positive sense of the term, it is sort of like a specialized version of Chad GPT and it works really well in the office environment is also what used to be called Bing copilot is to me the tool that I tend to use if I’m using AI that goes out to the internet, which I don’t like that.
And we could do a whole podcast on why I think the combination of AI and searching the internet is less than ideal in a lot of situations. And that’s also where people say, if I’m going to do that search, I maybe want something more fine tuned like a perplexity. Yeah, so there’s pros and cons of copilot. I would say that if you’re not allowed to use other tools, this is the one you’re most likely to have available and you can do all the things that, so again, it’s another way to learn prompting do other things. It is just going to be inside the office environment. So those are my thoughts on copilot.
Tom Mighell:
What about Claude Tom? Claude is the only one that I don’t have a subscription to, so I know less about this. I have no personal experience with Claude. What I hear and what I understand is it seems like based on what I hear about Claude is that it’s like the brainiest of all of them or the snoots, it’s known for its detailed explanations. It does really good at summarization. It has apparently a longer what they call a context capability. You can put a lot of context in there. It’s great for sophisticated writing if you want to do good writing. It apparently does a better job than some of the other tools. It does good deep reading. It has detailed document analysis there. If there was any negatives that in my research I’ve learned about Claude is that it’s overly verbose. I don’t know that that’s necessarily a negative, but obviously if it goes on and on and you’re getting too much, then maybe that is, but I guess it makes up for its verbosity. Is that the right word with accuracy that the accuracy is generally a little better than some of the earlier GPT variance. But I don’t know, Dennis, you’ve used Claude a lot more than I have. Did I get any of that right? Is that accurate or how would you think about
Dennis Kennedy:
Claude? Yeah, I mean sort of some of the things that you were talking about there can definitely be handled through prompting to correct some of those things and to get your results closer to what you want. I a hot and cold on Claude. I mean I do pay for it and it was the one I was thinking about dropping recently, but they have a new version that became a little more interesting to me, so I haven’t dropped it yet. I feel like it’s the one I have to fight the most with when I am prompting because it prefers simpler prompts and simpler output and shorter output. And so it’s harder to have it do several things at once. And it does have a tendency to ask you questions, do you want me to do this? And you say yes. And then you feel like it’s saying, do you really want me to do this?
I just found it really frustrating. It is the one AI tool that I got to spontaneously make a joke back at me because I complained about how it kept asking me to do things. And so I may post about that. It was kind of funny how I did it. So you see people who are saying, I’m looking for a second one to try because I hear about all these other tools and I’m just using chat GPT. Maybe it is a fomo, maybe I need another one. So people use Claude a lot. I think it’s a decent choice as a second tool. And again, I’m looking to say what tool makes sense for me to learn prompting what I have to learn to make this stuff work? And so Claude I think is fine. So if you say I’m concerned about being the open AI world, that’s a lot of people made the Claude Switch for that.
Some people say this odd thing to me, they say Claude is a better writer, and that one’s weird to me because I prompt the output that I want. And so I’m not really relying on its default approach. So that’s an interesting thing to me. But there’s some parts about it where I do feel like I fight it where I feel it’s trying to make me do things instead of doing what and do things in its way rather than me doing what I tell it to. And then as I said, Tom, I’ve become a big fan of Google Gemini advanced recently, but I’m really curious about hearing your thoughts on that.
Tom Mighell:
Well, so what makes Google Gemini? Well, I guess I want to start first by saying that Google originally had a really bad disadvantage because like you said, they stumbled pretty badly out of the gate and they really didn’t introduce it. They were slow and late. And when they actually did come and debut, it was not ideal. It made mistakes and it really didn’t give people confidence that it was a good tool. But it has caught up a lot since then. And I think there are some who would say that it may be one of the best ones that are out there. What I like about it is as an Android user is the fact that it’s multimodal, which means I can access it anywhere. It’s built into my pixel phone. So that now instead of just having what used to be the Google assistant, I now have the Gemini assistant.
So when I talk to my phone and ask for things, I’m actually talking to Gemini, which is nice. It’s a different intelligence. It is a different set of things. I am having to learn what it can and can’t do. So I think that Gemini takes advantage of Google strengths in a good way. I think that’s what I like about it. I think that what is especially intriguing to me, and we’ve talked about this multiple times on the podcast, is how its use through Notebook LM and being able to, and I know we’ll talk about this a little bit more later about how you can use Notebook LM against your own content and basically create sort of do, and Dennis, tell me if I’m wrong if I’m using this in the right context, but sort of your own retrieval, augmented generative copy capabilities because basically turning an AI against your own materials.
Dennis Kennedy:
Exactly.
Tom Mighell:
I like that. And that’s something that none of the other tools are doing in the same way or with the same success rate. So those are my major things. I don’t use it that often just for daily stuff. I know you’re doing that a lot more often these days. So maybe tell us a little bit about your experiences with Google.
Dennis Kennedy:
Well, what I like about Google is it kind of came out of nowhere because it did stumble really badly. And then in this Gemini advanced two world, it was shocking to me how well it was doing and I kind of lucked into it because I paid for Notebook lm. And then so I started to experiment and it was like the input window just seemed enormous to me, and I haven’t checked any of the details on it. This is sort of my feeling for it that seemed enormous. The output was not as limited. I could specify more things and then they roll out a number of different versions. And so you got to see not just the sort of generative AI that we traditionally think about, but the sort of deep thinking approaches and other things like that, the show your work kind of approaches that have come out.
So now if you are in the Google Gemini 2.5 world, it is a different world of AI and the results that you can expect than what even the sort of GPT-4 level a few months ago. It is really sort of astounding what you can get out of these tools. And I’ve really sort of turned to the Google Gemini advance for a lot of things because I have so much control over the output and I can kind of really tweak it in ways that I want. That downside I would say of the Google tools is that there are a lot of options and they have similar names. It’s a little hard to figure out what tool you want to use, but I typically just go to the most advanced experimental form and see what’s happening at the very high end. There are pros and cons of that obviously, but that’s what I like to do and that’s what makes it fun to try this tool.
But some of the output that I’ve done with Google Gemini Advance, I think the thing that I would use that I might say like, oh, and if you’re using generative AI in the normal tools, say at the end of say last November-ish, I could say, if you’re writing a paper for law school or something like that, I would say, yeah, you’re, it is probably going to give you B minus type work. I would say these days, if you’re good at prompting, I think you’re going to get significantly better results from that. But I still don’t think you want to do that. But the analytics you can do the other things that can help you with. I think that if you’ve not used the very high end tools like the Google Gemini, you’re going to be shocked at what you’re seeing the ais do now. So I guess, Tom, before we go to the break, I think what’s made, I’ve said this about Google Gemini where I see it a lot, but it’s making, comparing tools so difficult right now is there are so many flavors of each of the tools and we’ll dig into that after the break.
Tom Mighell:
That’s right. Let’s take a quick break from a word from our sponsors and we’ll be right back.
Depositions. Offer key moments to turn a case in your favor if you can find them. Introducing Verit legal Visor, the real-time verbal intelligence platform for litigators. Transform discovery with real-time inconsistency, detection, deposition summaries and live translation. Receive alerts for discrepancies in a witness’s testimony or inconsistencies with past testimony or case documents. Legal advisor leads no word unturned, giving you the critical edge to win all from the global leader in real time. Legal transcription. Visit verit.ai today to schedule a demo
Dennis Kennedy:
And we are back. Tom, let’s talk about the flavors or models of the AI tools. As I mentioned before, the break in Gemini Advanced, I have as the time I wrote the script, I have five choices of models in chat, GPTI have eight, and in Claude I have six. Some tools will also do image and video generation. So that is probably an overwhelming number of choices even within that big choice of whether you’re using Claude Chat, GPT, Gemini, copilot. So we need to think, and I think what could be useful in this podcast is how do you sort through that? And we’re starting to see a number of ways that people are looking at to try to sort through those things. And so the one that I’m seeing more and more of is the notion of benchmarks and whether some sort of benchmarking can help us. So I have very mixed opinions about that, but Tom, I want to see what you think about benchmarking AI tools and whether that’s something that we as users might rely on.
Tom Mighell:
Well, I would say first you’re right. I said I want to simplify this and only do four LLMs, but considering how many flavors we are, we’ve now expanded that essentially to 20. I mean if you add them all together, we have 20 different versions and it’s no longer enough to say Claude versus GT four, the comparison should be Claude three Opus versus GPT, turbo versus Gemini Advanced and what does that mean? And so I have mixed feelings about benchmarks as well because I think benchmarks, one, they’re measurable. They offer measurable ways to compare performance across specific tasks. So you can see how you can do things. So factual accuracy, you can choose which ones are factually better, how reasoning is better, how well it follows your instructions. So that’s useful to be able to get that. So that’s a good way I think to be able to objectively look at these models, especially if you have very clearly defined tasks.
Here’s what I know I want to do and here’s what I have. But I think that some of the limitations that benchmarks might have is that they may oversimplifying the real world use cases. So your Mighellage may vary I guess is probably the way to think about it is that especially in a nuanced domain like the law, and especially in an area where we’re talking about the law, it may not reflect, these benchmarks may not reflect everyday complexities where things may be different. You may be getting different answers. I think that also higher scores or higher performance on benchmarks doesn’t equal better user experience or better satisfaction. Just because the numbers are better doesn’t all automatically mean you’re going to have a better experience or productivity gain from that. And then it’s really those two reasons that make me think that. I think benchmarks are probably a useful factor to consider, but where I really think that the bottom line here should be is that you really need to think what specific flavor or which one of the tools matches your workflow. Do you want a rapid response? Are you looking for creative content generation? You have high context tasks that you want to achieve there. Being able to say, here’s what I want to get out of it, is really going to drive, I think more of your decision making than what the benchmarks are going to tell you.
Dennis Kennedy:
Yeah, I really struggle with benchmarks because there’s sort of three reasons. And so one is that people are looking for some sort of objective measure and that I don’t know how relevant it is. So if one version could do some task in half a second and it used to take two seconds, I don’t know that I actually care about that. So is the benchmark really meaningful to me? I think the benchmark that becomes really interesting to me is how does this compare to a human doing the same thing? And we’re starting to see a little bit in the way of those benchmarkings. I think that’s interesting. And then because my focus in AI is so dominated by prompting that benchmarking is I need to know, well, how was it prompted? Because the prompt is going to change the results. So I struggle with benchmarks. I know that we’re trying to figure out some way to compare, and it’s nice to have some notion of being objective, but I’m not sure that benchmarks these days help us that much unless we’re engineers deeply into that.
So benchmarks I don’t think really do it. I think that although there were other things we could talk about deep seek and open source LLMs like LAMA and Misra and all these other things, but I think that we’re looking for frameworks and I have reached Tom as, and this won’t surprise you, this point where I feel, I actually think it’s true, I’m now spending more for AI subscriptions than I am for cable tv, which is kind of an interesting place to be. But I think what’s really helpful to people is to say, and I did this in my class the other day, is to say, how would you pick a tool if we have something that’s changing so rapidly that we’re not sure about the benchmarks, we’re not sure about what it does exactly, how we would compare those results. And we have a couple of different variables in there. The biggest one for me being prompting. So Tom, I’m ready to hear what you have to say about picking an AI tool.
Tom Mighell:
So I think that there are, I kind of narrowed it down to five considerations, five things that you need to think about. I think the first thing is what are your goals? What are your core goals about what you’re sort of to come back to our usual thing, what are you hiring the tool to do? What do you want to have happen? Do you mainly do legal research? Do you do document drafting? Do you do summarization? What do you do? And could be all of the above, but figure out your requirements first. Second thing is ease of adoption. How steep is the learning curve for the tool that you’re using? Now for most of these, there’s not a tremendous learning curve that’s different between each one of them. So this may be one of the easier one, but also think about how well that tool can integrate into the workflows or the apps you already use.
Are you using Google Workspace? Are you using Microsoft Office? That may lean you towards the tools that are available there? It may not, but think about that. How easy is it for you to integrate a tool into what you’re already doing? Third has to be budget considerations. Dennis is right, you should spring for one of the tools and pay $20 a month. I’m not comfortable spending more than I pay on cable for all of the tools, so I’m never going to do that. But you need to think about balance between the cost, the functionality, the time saved. We think you should be spending some money. And frankly, what you should do is, I’ll jump down to my last one, which is that trial and error is okay. You don’t have to focus on one right now. You can focus on one or you can try ’em all out there.
The tools are things that you can subscribe to and then end your subscription immediately and start up with, pick up with another one. So do some experimentation, try all of them and see what you like. The last piece that I will say is understand clearly where your data goes. Understand security and privacy. That’s going to be because we are lawyers and that needs to be, what we need to think about is legal professionals dealing with potentially sensitive information. You need to understand who has access, what happens to your data, be aware of all of that before you make a decision on a tool. Those are to me, the five main criteria that you need to think about. And hopefully that leads you to a good answer on one of these tools.
Dennis Kennedy:
Do you have a good acronym for that? Tom
Tom Mighell:
Sebs, I think is what it turned out to be. SEBS or est? I don’t know. No, I don’t have a good acronym for that.
Dennis Kennedy:
So I guess what you’re saying is you don’t have
Tom Mighell:
A good acronym. I don’t have a good acronym. That is correct.
Dennis Kennedy:
So a couple of things that I look at. So we’re in a period where there’s too many choices, so we just need to simplify the choices. So to me it’s like spend $20 chat, GPT, be done with it, learn to use it, experiment with it, that sort of thing. If you have some specific needs, then I would explore some other things. So if you’re just interested in AI and you say, I would like to try two things and compare them, you might do that. So I really harp on the $20 a month in paying for this stuff because if this is the most important technology development of our time, then it’s going to be weird to look back and say, yeah, this is the most important thing that happened in technology in my lifetime, and I wouldn’t even spend $20 on it. So I always do keep that in mind.
So a couple of things. So just pick one. I think there’s an easy choice for that. Then what I tell people is look at low risk stuff. I mean, people are using AI all the time. Think about other places where AI has kind of come into your vision or it’s used on things and people are always going like, oh, I’m looking for a way to replace this. I’m looking for a way to write the perfect brief with the AI on the first try. I’m like, think about something. Look at planning your vacation. Look at stuff that’s easy. Look at something you have a hobby that you’re interested in. That’s the place to play with it and look for these low risk things. And then at Vanderbilt, I went to this women in AI summit and I came back with this note to myself that said, we need to approach AI not with a sense of fear, but with a sense of wonder, a sense of experimentation and a sense of fun.
I think that you’re looking for ways that you can try AI that would be fun. And where you say like, well, I shouldn’t be afraid of this because it’s a low enough risk. And then the point that I’ve made a couple times is I think you want a tool that allows you to learn prompting and learn it well. And so that again, to me, I recommend to people is like, look for something that’s a hobby or an area of interest where you’re not trying to say, I’m trying to figure out how to replace some complex thing that I do in my work. Let me learn it on something that would be fun, fun to learn. So those are my thoughts, Tom.
Tom Mighell:
And with that, you are on your own. Go and find a tool. Let us know what you’re using. We’d love to hear what you’re using. Maybe we’ll do an informal survey sometime and figure out what our listeners happen to use. We are running way behind on our time, so we have got to take, but we’ve got to take a quick break and then we’ll be back to quickly wrap up the rest of the show.
And now let’s get back to the Kennedy Mighell report. I’m Tom Mighell and I’m Dennis
Dennis Kennedy:
Kennedy, and we wanted to remind you to share the podcast with a friend or two, especially this one that helps us out. In this segment, we turn again to a generative AI tool to act as our stand-in for you, our audience, and create a question for us to answer. So this time I asked Chad GPT to give us a question appropriate for this episode. And here’s the question. Can Tom guess correctly what AI tools Dennis is currently using most and why? Tom, what say you?
Tom Mighell:
Okay, first, I say that’s an absurd question to make me try to guess what you’re using. I’m never going to be able to guess, although I will say in the pre-show and when lots of things that you talked about in the pre-show gave me clues. Lots of things you said here gave me clues, but before I had any of those clues, I decided I was going to use my friend chat GPT and say, here’s my friend Dennis. Here’s what I know. Tell me what you think so we can make a decision. So chat, GPT told me to say the following, Dennis, you experiment with everything, but if I had to guess your current favorites, I’d say you’re currently leaning heavily on Claude three because of your love of thoughtful, detailed writing. You’ve been notably excited about Google Gemini Advanced and Notebook LM for teaching and organizing your ideas and research and knowing you, you’ve likely got your fingers into some open source tools. As for chat, GPT, I guess it’s still around, but you might’ve moved on for it a bit lately, and you’re not spending much time with copilot at all. Based on what we’ve talked about today, I would alter that slightly and I would say you’re probably spending more time with Gemini than any of the others right now, less with Claude than I expected. And you’re still probably using chat GPT slightly, again, not using copilot at all. So how well did my friend chat GPT do at getting those guesses?
Dennis Kennedy:
If you swapped out Google, Gemini and Claude, you’d be exactly right. I did tell you that I almost stopped paying for Claude. So Google Gemini is the one that I turn to most, and there’s another reason for it to, and that’s the Google Notebook thing where I have a big new personal knowledge management project in Google LM that I’m really excited about that I haven’t really talked to people about yet. So that also gives me reasons to live more in the Google world, which I would say three months ago, Google was barely even on my radar. So I would say chat, GPT as a starting point did a really nice job with its reasoning. It’s just there were some other factors that didn’t take into account, but with the human in the loop and a few corrections, it was pretty darn good. So now it’s time for our parting shots, that one tip website, our observation you can use the second this podcast ends. Tom, take it away.
Tom Mighell:
Let me just say I’m going to have a couple of parting shots. My first parting shot is back on your last response, Dennis, which is I want to remind you all the times in the past 387 episodes where you gave me unmitigated grief by giving my life over to Google. And essentially you have now done the exact same thing. I just want to see how that whole relationship works out for you. But I’m glad to see that you’re embracing our Google overlords here with that. So that’s one parting shot. Second parting shot is around Google, also with Notebook lm, if you haven’t paid attention, we talked about in the past how you can feed Notebook, LM some information, and it will create a podcast of that data. It now offers an interactive mode where you can insert yourself into the podcast that the two people are having a conversation.
You can click Answer or I forget what the number word is. You can click the button and when you click the button, the two speakers go, oh, somebody wants to ask us a question. And you can talk and have a conversation with them and ask them questions. It’s a little bit like talking to the chat version of Chat GPT or Perplexity, but it’s interesting that you can now talk to them about your data. So go give it a shot. I think it’s interesting. My other parting shot is if you’re not using your Outlook calendar and Microsoft Teams, you should be doing that. You should be setting up meetings in teams using your Outlook calendar. And what’s especially nice about it is I have for my delivery team, we have a meeting every two weeks to talk about delivery items. I’ve set it up automatically so that when the meeting opens, copilot is automatically transcribing and sitting in on the meeting. It happens automatically. And I’m able to create an agenda and some meeting notes and tasks that automatically open using Microsoft Loop so everyone can add into the notes, people can add tasks as they get added. The agenda can switch around if people want to move things and talk about things. It’s a super way to keep track of meetings now in teams. So if you haven’t paid attention to it, look at it. I think it’s a really great upgrade in Teams. Dennis.
Dennis Kennedy:
So Tom, I have one that you could actually use in your 40 tips from the last 40 years thing that you’re doing at Tech Show is I looked at the Notebook lm and this project that I’m doing, the way that Notebook LM works is that you typically will want to have once PDFs basically. So I was thinking about that. And so I was looking at Adobe Acrobat and it was a blast from the past time. I just said, wait, what if I go to things like my blog and other things to get and the data into Notebook lm, what might be the way to do it? And I was like, wait, I just create a PDF from the webpage, and I type in the URL and it just grabs it all and throws it into PDFI save that PDFI uploaded it to Google Notebook and it becomes part of my personalized AI in there.
And so that simple tool is kind of a blast from the past, is actually turning out to be an important engine to pull information, my own personal stuff into the Google Notebook. And then the other thing that I always, I like to give as a parting shot is new issue of my monthly newsletter on personal quarter layoff sites called Personal Strategy Compass. It’s on Substack. I love it. I just finished a personal quarterly offsite here at the end of March, and it was, I think the personal quarterly offsite is probably the most helpful thing I’ve done and the most useful thing I’ve ever created. And I love getting the chance to write about it and share this approach with other people.
Tom Mighell:
And so that wraps it up for this longer than usual episode of the Kennedy Mall report. Thanks for hanging in there with us. Thanks for joining us on the podcast. You can find show notes for this episode on the Legal Talk Networks page for our show. You can find all of our previous podcasts along with transcripts on the Legal to Network site, and you can subscribe to our podcast there as well, or you can subscribe to us in your favorite podcast app. If you’d like to get in touch with us, remember, you can always find us on LinkedIn or if you’d like, leave us a voicemail. We still want those questions, folks. That number is (720) 441-6820. Maybe you’ll call us and tell us what AI tools you’re using and why you’re using them. So until the next podcast, I’m Tom Mighell.
Dennis Kennedy:
And I’m Dennis Kennedy, and you’ve been listening to the Kennedy Mighell Report, a podcast on legal technology with an internet focus. If you like what you heard today, please rate us an Apple podcast. And as always, a big thank you to the Legal Talk Network team for producing and distributing this podcast. We’ll see you next time for another episode of the Kennedy Mighell Report on the Legal Talk Network.
Announcer:
Thanks for listening to the Kennedy Mighell report. Check out Dennis and Tom’s book, the Lawyer’s Guide to Collaboration Tools and Technologies, smart Ways to Work Together from A Books or Amazon. And join us every other week for another edition of the Kennedy Mighell Report, only on the Legal Talk Network.
Notify me when there’s a new episode!
![]() |
Kennedy-Mighell Report |
Dennis Kennedy and Tom Mighell talk the latest technology to improve services, client interactions, and workflow.