How AI Search Is Reshaping Law Firm SEO — and What To Do About It with Michael Buckbee

Michael Buckbee (00:01.676)
Hi, I’m Michael Buckbee, the co-founder of Knowatoa. We’re a service that helps websites and brands rank better in AI search across all the different services and all the new weird ways that search is changing these days.

Karin Conroy (00:15.93)
Michael, thank you so much for being here. I was just telling you that I am very excited about this because for a lot of reasons, AI search is the thing that is on everybody’s mind and the topic of the day, AI in general. And AI search specifically, we’re gonna talk about what that is exactly, but AI search specifically is…

kind of a mystery a little bit to even those of us who are in the industry, we’re kind of watching it unfold at the moment by moment. So we’re gonna talk about how that relates to law firms, what it is, what you need to be paying attention to, all the things that are kind of happening in the moment, and then maybe a little bit of like what might be on the horizon. So the title for today’s show is How AI Search is Reshaping Law Firm SEO and What to Do About It.

So let’s dig in. Let’s start first, like with the basics basics. Like what, what is AI search?

Michael Buckbee (01:20.606)
so I think the best place to start is to pretend like chat GPT and Claude and all these new services don’t exist. Like, and you know, just pretend they don’t exist and pretend we’re only dealing with Google. so Google has made some decisions and the same way they’ve made decisions for the last 20 years about how they present search results, how much of them are ads, how much is like called a zero click, like a summary of information and.

Karin Conroy (01:33.965)
Okay.

Michael Buckbee (01:49.186)
they have made a decision to go really, really hard at something called AI overviews. And so AI overviews are a summary of the top set of results and how many results and which ones and all sorts of stuff is changing constantly. But it’s a summary of the top sort of results. And to date, this is mostly affecting informational search queries. So when people make searches, there’s all kinds of search like, hey,

Karin Conroy (01:55.117)
Okay.

Karin Conroy (02:06.895)
Yeah.

Michael Buckbee (02:18.318)
I want to buy something or I’m trying to get an answer to a question. And predominantly these AI overviews are for the answers to questions. And the last stat I saw was that for logged in users, because you have to be logged into Google to get these. it’s about 70 % of their queries have an AI overview.

Karin Conroy (02:33.443)
interesting point. Okay. All right. So for logged in users.

Karin Conroy (02:41.498)
Oh, okay. That’s fascinating. So for, for those of you who, um, aren’t following along with the, the title of the AI overview, this is the thing you see before you see all your results. So you see the little kind of blah, blah, blah, blah of the AI like kind of taking your question in and kind of spinning and like processing, processing. And then all of a sudden there’s that answer at the top that

Michael Buckbee (02:55.213)
Mm-hmm.

Karin Conroy (03:09.858)
In my experience, when I’m asking especially basic stuff like where is there a sushi restaurant near me? It’s giving me some solid answers that I do not need to click on. I’ve got my answer already. And so I just wanted to underline the point that you made a minute ago about these click-free answers. And the reason this is important is because when it comes to SEO,

Michael Buckbee (03:15.598)
Mm-hmm.

Michael Buckbee (03:31.884)
Yeah.

Karin Conroy (03:36.908)
you are counting on clicks. That’s how you measure everything, right? And so we’re in a different world now where people are getting answers without having to click on a Google search result. And this is one of these points of differentiation where AI is coming in.

Michael Buckbee (03:40.279)
Yeah.

Michael Buckbee (03:52.942)
Yeah. And again, sticking just to Google, like, and to be clear, this is now as well. This is not a future thing. This is, this is what’s happening now. Google has a utility called Google search console, which if you have a website, you can register your website and you basically get sort of interesting metrics about search and how it affects. And what we’re seeing is, and if you want to know, Hey, are AI overviews affecting me?

Karin Conroy (04:01.891)
Right.

Karin Conroy (04:06.191)
Yes.

Michael Buckbee (04:19.074)
you go in and look and you see impressions, which is the number of times your site showed up in a search versus the number of click-throughs. And we’re seeing a big divergence in that across lots and lots of sites where their impressions are staying more or less the same, but their click-throughs are just falling off a cliff.

Karin Conroy (04:36.708)
Fascinating. All right. So let’s, okay. So that explains this AI search. There is a difference between, you you’re kind of focusing on this, what Google is doing. And there is a difference between that and then what people are seeing within chat GPT, Claude, all of the different AI platforms. But let’s talk about, should we jump into what to do next or should we talk more about your framework? What makes more sense like in the sequence of, of kind of understanding?

Michael Buckbee (04:47.469)
Mm-hmm.

Michael Buckbee (05:03.21)
I, you know, one of the really tough things right now is that there aren’t great terms that are defined for all these things. So I was actually going to be my suggestion, which is that we just take a second to define. So, when we think about this, we think in terms of traditional search volume, which is like how many searches a month for our, like, you know, divorced lawyer in Chicago. And then we think in terms of AI search volume, how much of that

Karin Conroy (05:11.138)
Yes. Okay.

Karin Conroy (05:25.475)
Yes.

Michael Buckbee (05:30.142)
an AI overview is being seen. And then we define a chat search volume, which is how much in tools like chat GPT and Claude and these different things are people looking for answers or for this type of information. And, you know, the same way that SEOs and marketing people sort of try to find, where are people at? And then to give them good information, you sort of need a little bit of a different strategy for each of those three different things.

Karin Conroy (05:35.267)
Okay.

Karin Conroy (05:57.612)
Okay, so what are you calling each of those three different things? Because there’s like the Google search, the Google search AI response that we were describing a minute ago, the chat, like within the chat bot. Do you have a title for each of those three sections?

Michael Buckbee (06:08.279)
Mm-hmm.

Michael Buckbee (06:14.828)
We just call them AI search volume and then chat search volume. And in the same way that so many of the searches now on Google don’t result in a click, it’s even more in chat GPT and the other services. And I think so much of the frustration and sort of concern we have is that people that are heavily using these services are having a much better experience where they’re like,

Karin Conroy (06:20.195)
Okay.

Karin Conroy (06:42.457)
Yeah.

Michael Buckbee (06:43.874)
I get a better answer out of chat GPT than I do going through this whole dance of like going to Google and looking at the top five pages and getting a little piece of each and then synthesizing that like judges. It just does it for me. So it’s very hard to compete with that. So yeah.

Karin Conroy (06:57.038)
Yeah, yeah.

Okay, all right, so what do we need to do next? How do we think about all of this stuff? Because there’s a lot going on, like you described. There’s the things going on in the chat, chat GPT, there’s the AI search. I mean, there’s just so many different things. How do we even think through a strategy on all of this?

Michael Buckbee (07:16.142)
Mm-hmm.

Michael Buckbee (07:22.326)
So we do have something called a biscuit framework and it’s an acronym for a bunch of different things to do and it’s agnostic. So it’s for all these different types of things. And it’s really to try to get people to take, Hey, here’s some basic, you know, technical things, make sure all these different new services and sites, and even just again, focusing just on Google, it used to be, there was a single Google bot that would index your site. And now there’s like two or three, there’s a Google extended, there’s other things.

Karin Conroy (07:25.315)
I love this.

Michael Buckbee (07:51.282)
And we found a lot of sites are accidentally or just unaware they’re blocking some of these. And if that’s the case, you’re going to be cut out of a lot of these. It’s a problem. It’s a problem. And so that’s one thing. And there’s a way you can check those manually. And then our service, Noatoa, has a built in way to check those for you too. It’s free if you want to try it.

Karin Conroy (07:57.593)
Yeah.

Karin Conroy (08:01.017)
It’s a problem. Yeah. Yeah. Okay.

Karin Conroy (08:12.89)
So that’s the B, like check to make sure you’re not blocking, right? Okay.

Michael Buckbee (08:16.11)
Yeah. And then, you know, I don’t know you want to go through all the acronyms because it’s a lot, but it is a longer word. But, but, you know, the other pieces are, you know, make sure that you’re indexed, make sure that, you know, not only is your information in there. And usually we talk about this in terms of brand. So we talk about, you know, just brand visibility and you can go into chat, GBT like, Hey,

Karin Conroy (08:20.984)
Okay, it’s kind of a longer word.

Karin Conroy (08:39.343)
Yeah.

Michael Buckbee (08:44.728)
do you know who my law firm is? And if it says like, yep, you operate in Houston and Texas, and you know, that’s great. If it’s accurate, it might not be accurate, or it might get you confused, or it might make up a law firm that kind of sounds like, which is, you know, a real thing that happens. And that sort of leads to the second point, which is we talk about brand knowledge. And so this is not usually something SEOs have been concerned about.

Karin Conroy (08:55.14)
Yeah.

Karin Conroy (08:59.866)
Yeah.

Karin Conroy (09:04.462)
Yeah.

Michael Buckbee (09:14.614)
It’s a lot more like internal business stuff, but just like, is the information out there accurate about your law firm? You happen to do, you know, divorce and commercial law. ChatGPT thinks you only do divorce law. Well, that’s problem. And that’s not exactly a search query. It’s more like this basic stuff. So this sort of brand knowledge now is a new thing. have to sort of monitor and think about and check. And then last.

Karin Conroy (09:19.161)
Yeah.

Karin Conroy (09:27.555)
Yeah.

Karin Conroy (09:32.601)
Right.

Karin Conroy (09:42.242)
Yeah, so, okay, so, sorry, sorry. Yeah, no, that, because then I want to come back and talk about how you address that.

Michael Buckbee (09:43.81)
Yeah, good.

Michael Buckbee (09:49.4)
Sure, and then the last was gonna be brand competition, which is like when you really get into like best divorce lawyer Chicago, that and those, and so there’s three distinct sort of areas of content that you need to create and put out there and monitor and track.

Karin Conroy (09:52.193)
Okay.

Karin Conroy (09:56.324)
Yeah.

Karin Conroy (10:06.936)
Yeah. Okay. So going back to the second one where you’re talking about, it’s sort of like the definition of your brand or the law firm. And like you said, this isn’t, traditional SEO type content, cause we’re not talking about like how to solve the problems. We’re not talking about the kind of questions people are asking. We’re talking about the, the definition or the description of your firm itself. How do you, let’s say you,

Michael Buckbee (10:33.751)
Mm-hmm.

Karin Conroy (10:36.302)
go into one of these platforms or chat GPT or Google or any of these options and it’s wrong. So what would your first recommendation be to address that and try to adjust it?

Michael Buckbee (10:50.831)
Sure. So let’s, take chat GPT. Cause I think that’s, that’s very popular and people know what that is. So chat GPT, there’s two pieces of this. One of which is they certainly have their own crawler. And so you do want to make sure they can crawl your website and go through it. And they are quite aggressive with that. If anything, like there’s these bots are wrecking some websites because they’re it’s not, they’re not trying to find out about you.

Karin Conroy (10:56.462)
Yeah, I love it.

Karin Conroy (11:17.018)
It’s like a bull in a china shop.

Michael Buckbee (11:19.202)
you’re actively blocking them. if you, yeah, so there’s a lot of things that maybe are almost too basic that you don’t have listed on your site, but you should certainly list them on your site. So that’s step one. And that’s mostly just because it’s a gimme. It’s your site, you control it. It’s very easy to update. The next thing is to look at where the data sources that chat GPT is feeding off of. So chat GPT,

Karin Conroy (11:36.674)
Okay. Yeah.

Michael Buckbee (11:48.226)
has a multi-billion dollar relationship with Microsoft. A portion of that is because Microsoft also owns LinkedIn. And a portion of that whole relationship is they get access to the LinkedIn feed. So what we’ve been encouraging people to do is to publish more what we call brand statement content. And it’s just like, hey, it’s 2025. we’re just really so proud of here’s our team of divorce lawyers that work in the Houston area.

Karin Conroy (11:51.428)
Okay.

Karin Conroy (12:05.91)
fascinating.

Karin Conroy (12:14.798)
Yeah. Yeah. Yeah.

Michael Buckbee (12:18.03)
And it’s a little silly to, I think, just state that so plainly, but where I think as a marketer, it is a good opportunity to reintroduce people and restate what you say and have things in different ways and to just, please go ahead.

Karin Conroy (12:31.759)
Yeah.

You know what this reminds me of is early SEO, like way back, I feel like this might’ve even been back when it was still HTML, when we used to stuff the footers of websites with like, especially for law firms, it was like every name of a city within like a 300 mile radius and the county and whatever. And there would be this footer that was like three paragraphs long. And sometimes it would be like,

Michael Buckbee (12:40.12)
Mm-hmm.

Michael Buckbee (12:47.446)
Yeah.

Michael Buckbee (12:56.034)
Yeah.

Karin Conroy (13:01.126)
black on black text and like, there was like very kind of questionable tactics, but it was like keyword stuffing. and you’re just sort of trying to aim the, in those times you were trying to aim the bot in a specific direction for those kinds of search. So this is a similar sort of, kind of directional thing where you’re saying, Hey, we want to capture these searches that are for Houston divorce lawyers. And so we’re just letting you know.

Michael Buckbee (13:03.693)
Yeah.

Karin Conroy (13:30.36)
We’re Houston divorce lawyers. Yeah.

Michael Buckbee (13:32.322)
Yeah. Yeah. Well, and you I, you’ve made a very keen insight, which is that I do think, AI search, an interesting part of it is chat GPT when they, they founded themselves and came out with their products were not thinking like, Hey, we’re going to disrupt Google. Like that wasn’t their plan. They were trying to make this more generalized tool and it just happens that it did that. And as part of that, there’s a lot of things that

Karin Conroy (13:51.128)
Yeah. Yeah.

Michael Buckbee (14:01.486)
as SEOs, I think we bring to this that don’t actually exist. All of the ranking factors of like site authority and EAT and all of this stuff, the only ranking factor we can say for sure matters across all the different AI search tools is repetition. more you’re repeating your message, the more often it comes up. And I often see the sort of like causation and correlation.

Karin Conroy (14:05.22)
Right.

Karin Conroy (14:20.168)
Michael Buckbee (14:29.856)
and people twisting those together where they’re like, what you need to do is get into, you know, Wikipedia because Wikipedia is referenced all the time. Well, it is because it’s Wikipedia. It is referenced all the time, but it’s not actually, well, not as a huge domain footprint is what I would say, you know, so it’s in all of these, it doesn’t actually count for more than if I went out of my personal LinkedIn and push like, Hey, Houston divorce lawyers.

Karin Conroy (14:37.837)
Yes!

Karin Conroy (14:43.504)
Because it has a huge domain authority.

Okay, okay.

Michael Buckbee (14:59.458)
Chat GPT doesn’t make a distinction between those when it’s pulling the information in. It’s just two different, it’s plus one for each of those, if that makes sense. And that’s a weird mental model for SEOs, which is like, got a link from the New York Times, I got a link from Wired Magazine or a law journal. None of that exists. So it’s a different thing.

Karin Conroy (15:03.534)
Okay.

Karin Conroy (15:08.26)
Okay, okay. Yes.

Karin Conroy (15:27.566)
All right, that’s fascinating. While you were talking, that was like the next thing I was gonna ask you about is EEAT. So I love that you brought that up, because it makes for a smoother transition. But so you were saying that there are all these things that we’re looking at in terms of kind of models of how to get your content out there and things that these, know, chat GPT and these different.

What are you calling the actual search? Are they search bots that are coming to your site from chat GPT or? Not is that not right?

Michael Buckbee (15:58.946)
Yeah, there’s their bots and their bots in the same way that like Google bot is a bot that, you know, reads your site and then indexes it for later. There is a slight distinction between interactive and sort of indexing bots. And so there are some bots that come and like here’s an example. Google has a tool called Deep Research, which if you put in something like, hey, find me, you know,

Karin Conroy (16:02.745)
Okay.

Yeah.

Karin Conroy (16:15.746)
Okay.

Michael Buckbee (16:28.622)
all of the legal firms in Virginia Beach where I happen to live, it will go out and it will Google for you and then take those results and like chew them up and then make a nice report for you. And when you run that, it will actually give you updates as it goes. Like, hey, checking this site, hey, checking this site. And it is a bot at that moment reading that website, pulling the HTML down, interpreting it, rejecting some things, coming back.

Karin Conroy (16:40.079)
Okay.

Michael Buckbee (16:56.234)
And then at the end, gives you here’s your list of 20. Here’s the three I think are good, you know, which is a little terrifying because it is it is a step removed from the human even putting even seeing the site, you know.

Karin Conroy (17:01.026)
Yeah, that’s amazing.

Karin Conroy (17:06.914)
Yes. Yeah, right. Exactly. Okay. So you’ve got these bots, you’ve got Chetch.pt, they’re coming out to your sites, they’re making this analysis. And then there’s these formats like EAT that people are saying that this is sort of the approach you need to think of. But a minute ago, you were saying that really the best thing that you have found in terms of approach or strategy or plan is the repetition.

Michael Buckbee (17:16.974)
Mm-hmm.

Karin Conroy (17:35.012)
So let’s talk for a minute about what EAT is and why you are saying repetition, and maybe I’m understanding this incorrectly, but are you saying that repetition is proving to be better than EAT or just different?

Michael Buckbee (17:49.518)
Google has been abused and manipulated by, well.

Karin Conroy (18:00.196)
never heard people talking such an empathetic, sad way about Google. That’s awesome.

Michael Buckbee (18:05.55)
It’s a tough problem. Well, it is a tough problem. I think, know, setting aside users, something I hear from a lot of just normal people who are not SEOs, who are not developers is, man, anytime I search for anything, I get an SEO article. And what they mean is it’s an article that has a lot of extra information in it for Google that is not pertinent to me. And the classic is the recipes, which is like,

Karin Conroy (18:07.448)
Yeah.

Karin Conroy (18:27.258)
Yeah.

Karin Conroy (18:33.518)
Yeah. Seriously, give me the ingredients. I’m scrolling for five minutes. I don’t need the story. I don’t need all of this. Yeah. And the ads and the pop-ups, it’s so, that is such a good example.

Michael Buckbee (18:33.888)
I just want the apple pie recipe and there’s a whole thing about yeah. Yeah. The grad school, the crisp fall air, like all of that.

Michael Buckbee (18:49.838)
Right. So, and that’s what, that’s what people mean about SEO and like EAT was expertise, experience, authority and trust. And it was a way for Google to try to differentiate these things and like, it’s health information. It should be coming from a doctor. This person, seem like an actual doctor. Like we’re going to do that versus like me. If I said, you know, go stand in a bunch of sardines and it’ll cure your issues. You know, I did not have

Karin Conroy (18:53.913)
Yeah.

Karin Conroy (19:00.793)
Yes.

Karin Conroy (19:16.346)
Yeah, so then the doctor’s results get pushed higher because it’s noticing some level of authority or recognition of that results expertise. so Google has a way of evaluating that. That’s what the EEAT thing is.

Michael Buckbee (19:19.598)
Thanks

Michael Buckbee (19:28.11)
Hmm?

Yeah. And there was even some recent clarifying statements around that, that that was really only for your money or your life queries, which were the very important health queries, the other things, and that Mike’s video game review site doesn’t really meet that bar.

Karin Conroy (19:40.034)
Okay. Yeah.

Karin Conroy (19:48.122)
Not for you and I, but for somebody it might.

Michael Buckbee (19:52.366)
Fair enough, yeah. yeah. so, but that’s one of like a hundred different things that as SEOs we’ve always, we’ve been trained sort of to look for like, does it have an H1 on the thing? Does it contain the same words as what’s in the query? Like all of that stuff. But you know, how the actual LLMs are built is just literally the technical term was a bag of words that was like the algorithm. Like.

Karin Conroy (20:17.402)
Yeah.

Michael Buckbee (20:19.18)
We just got a big bag of words and we threw all the words in and we shook it. And then we got these patterns out of it and things. And if you’ve ever used anything like Markdown or if you just like copied, like you went to your law firm’s homepage and you just copied all the text off it and pasted it into a Word document, that’s kind of what it looks like to an LLM. And so, you know, what we’re recommending is, you know,

Karin Conroy (20:41.155)
Yeah.

Michael Buckbee (20:47.086)
Try to do a really good job of actually answering all the questions. Repeat those questions and answer them in as many different ways. And then syndicate and push that information out to as many different places as possible. Don’t worry about replicating the same thing multiple times, which is always a big SEO thing. Yeah.

Karin Conroy (21:05.316)
Well, that was my next question. I heard, I saw this whole article about making sure that, especially when it comes to your company title and like tagline, as well as your title and like, you know, I guess if you have sort of a tagline behind your title, that you use the exact same phrase over and over because you’re teaching these LLMs. So for example, if your company’s like my company’s

Michael Buckbee (21:21.334)
Mm.

Michael Buckbee (21:30.93)
yeah.

Karin Conroy (21:34.234)
tagline is creative representation for the legal industry. So if I use those exact words after the name of my company over and over and over and not kind of do variations because in the SEO of history, it was do all the variations, put the different last part of the word, put plural, put all the different possible variations. And now it’s like, no, don’t vary, stay.

Michael Buckbee (21:37.699)
Mm-hmm.

Michael Buckbee (21:57.101)
Yeah.

Karin Conroy (22:02.234)
So what have you seen on that? Is that kind of legit?

Michael Buckbee (22:10.105)
Well, I think it is legit, because I do think it’s the ranking factors repetition. So the more you repeat that, I would not be overly worried about the exact words of it, because it does go through a tokenization. I don’t know if you’ve heard that term. But you can actually look this up. There’s free tools online to let you do this. If you really want to understand, how does Chet GPT actually

Karin Conroy (22:22.616)
Okay.

Karin Conroy (22:26.071)
No.

Michael Buckbee (22:36.002)
do with this text and it’s you put it through this tokenizer and it just takes the words and it makes them into little chunks of words. And as part of that, it gives them numbers. And then it’s these long sequences and numbers. And, you know, if you’ve heard like, all these things, they aren’t really smart. They’re just like guessing at what the next word is. This is how they build up that, that those statistics to guess what the next word is. So a lot of times it is emitting stuff like pluralization and other pieces and things.

Karin Conroy (22:42.552)
Okay.

Karin Conroy (22:47.716)
Cool.

Karin Conroy (22:58.266)
Got it. That’s cool.

Michael Buckbee (23:06.05)
But more for, you know, here’s a slightly different path into this, which is, you know, we’re talking, we’ve been talking about web search primarily here, but there’s other types of searches. Like there’s definitely local search, which has a lot of overlap with the legal community and like e-commerce search. And those are search types that are not so dependent upon you having a website.

Karin Conroy (23:25.154)
Yes. Yep.

Michael Buckbee (23:32.28)
They’re a lot more dependent upon citations, which is just like you list the name and address of the company and that’s on the local newspaper and that’s on all sorts of stuff. And you just want to get more and more of those because, you know, it’s a ranking factor. And I feel like Google looks at local and they look at their e-commerce and they look at how much money they make from that. And they kind of go, why are we giving so much traffic to all these websites when we could, you know, and it’s

Karin Conroy (23:32.281)
Yeah.

Karin Conroy (23:44.804)
Yep. Yeah.

Karin Conroy (23:58.616)
Right, right. And it’s a control thing too, because once people land on the website, Google doesn’t have as much control over the messaging and the direction and whatever. Now, you know, those of us who own the website, now it’s kind of in our lap. So I feel like Google’s like, no, no, no, we want to hold on to it.

Michael Buckbee (24:17.868)
Yeah. And, you know, so, you know, something I’ve had to say to a lot of SEOs is like, Google’s problems are not our problems. Like they can deal with them themselves. but there is a real change happening in how people do searches. And in particular, the search journey where, if someone’s listening to this and they haven’t tried chat GPT, I would really recommend you just get on and play with it. Cause at the end of the day, it’s just typing into a different box than Google, but

Karin Conroy (24:25.412)
Yeah.

Yeah, right.

Karin Conroy (24:42.521)
Yes.

Karin Conroy (24:45.88)
Yes. Yeah. And you kind of like, this is the way the world is going. Like, come on, let’s get with the program. Let’s, you know, like this is what everybody’s talking about. You’re missing out.

Michael Buckbee (24:47.65)
The experience.

Michael Buckbee (24:52.174)
Yeah.

Yeah. So, you know, and I think as part of that, so much of what used to be, the strategy that, I’m going to stand up a website. want to attract people that again, I’m going to pull out divorce, you know, like maybe a piece of content I’m going to write as part of my strategies. Hey, here’s the top 10 things you should do before hiring a divorce, you know, lawyer, you know, classic sort of thing. Well,

Karin Conroy (25:09.795)
Yeah.

Karin Conroy (25:17.646)
Yep. Yep.

Michael Buckbee (25:21.354)
Now that’s something that chat GPT or a Google AI overview may just give you directly. There’s a real interesting case study that we saw with HubSpot. People in the legal industry may not be too familiar with HubSpot, they’re like a real… Okay, so HubSpot, they ranked incredibly well for tons and tons of different SEO terms. Often they were not…

Karin Conroy (25:26.585)
Yeah.

Karin Conroy (25:36.246)
Noah, a lot of my clients use HubSpot, yeah.

Michael Buckbee (25:47.874)
really related to their core business of like sending emails and stuff. One of which was resignation letter. And so, you know, if you typed in resignation letter, your intent is I would like to get a resignation letter template that I can then put my name in and my boss’s name in and I will email this to them. Well, and I’m done. And if you landed on HubSpot’s page, what you got was the recipe thing, which is like, I asked all my friends about their times, know.

Karin Conroy (25:49.498)
Yeah.

Karin Conroy (26:04.718)
Yep. And I’m done. Yeah.

Karin Conroy (26:17.506)
And you’re scrolling and you’re looking, you’re scrolling.

Michael Buckbee (26:18.058)
And and your scroll and and and they and they had EAT because they had a resignation expert and you know and quotes. And then you went to a separate page. I’m not kidding. You went to a separate page that was in the landing page where it had a progressive form that you would fill out before you’d get the template. And the overview for that is literally here’s like an embedded word doc of the resignation letter.

and you click and you can edit it in Google Docs or you can click and it will become a template in Gmail. And I bring that up because there is no way for anybody to compete with that. you can’t, can’t EAT, you can’t make your titles better. can’t, you can’t do any of that. So, you know, you need to change your strategy, which is to move more towards the super competitive terms that, you know,

Karin Conroy (27:00.344)
No, yeah.

Karin Conroy (27:06.51)
Yeah. Yeah.

Michael Buckbee (27:17.004)
what is the best law firm in X, you need to absolutely win those. Cause you can’t win the informational ones anymore cause they’re going away.

Karin Conroy (27:23.17)
Right. Yeah. Well, and it’s almost like Google was taking that HubSpot sort of lead magnet and grabbing it and pulling it out from behind the curtain. So that leads to the next thing that we wanted to talk about, which is content ownership and all the legal challenges. And this is such a big, big topic about, I’ve seen lots of creatives, colleagues,

Michael Buckbee (27:41.336)
Sure.

Karin Conroy (27:52.25)
who will not use these LLMs because they think they’re thieves, basically. So let’s talk about that for a minute. Where is the line and where can law firms safely use this without feeling like they’re… How do you know if you’re taking content from some… You start to ask it some questions and how do you know if it’s not taking content from your competitor down the street?

Michael Buckbee (27:58.509)
Yeah.

Michael Buckbee (28:22.67)
Well, so I think there’s two aspects of this. There’s sort of the search part and the generative part. And so for the search part, it is very, you know, like, I guess the best way is to say it’s very reflective. Like you type it in and it says like, here’s the here’s the information and then you do with it what you want. But it is the case that all these systems were not are not really databases. They aren’t really retrieval systems.

there’s still creation systems for everything. So when you say like, Hey, give me this result. It is creating like a new report that looks like something it wants you to have, if that makes sense. So, yeah, so it’s giving you that. And I think ultimately it is really impossible to know all the sources that went into this. certainly it gives you some citations. gives you some reference, but

Karin Conroy (28:59.32)
Yeah. Yeah.

Karin Conroy (29:15.972)
I’ve seen that improve in the last version where I was asking it to help me plan a trip and it was giving me kind of a checklist of results for different things. And each item in the checklist had a link to the source of where it was finding that. But let’s say, which this is something I hear law firms doing all the time, which I feel like is on the other side of the risk boundary that I…

Michael Buckbee (29:27.662)
Mm-hmm.

Michael Buckbee (29:44.046)
Hmm?

Karin Conroy (29:45.062)
is let’s say you ask it to write you a blog post. How do you, is there a situation where this is truly fresh or is it pretty much always taking from some content that’s out there?

Michael Buckbee (30:00.716)
Well, let’s start with, let’s go back a couple of years to the, the, number one, like content technique that was pitched for years was the skyscraper technique. Are you familiar with this?

Karin Conroy (30:03.704)
Yeah, okay.

Karin Conroy (30:12.026)
it’s, I feel like that sounds vaguely familiar, but explain it to me.

Michael Buckbee (30:15.618)
Yeah, so the skyscraper technique was, hey, look at what your top like three competitors are doing for your thing and then just add 10 % to it and then publish that on your site. okay, technically you switched around the words a little and did stuff.

Karin Conroy (30:20.93)
Yeah, okay.

Karin Conroy (30:27.416)
Yeah. Yeah.

Right, yes. I will say, that totally took me back for a minute. I had a couple, not a couple, I had a while where I had a bunch of sites that were winning some design awards that I was like putting them in. And so they became, these particular websites became sort of popular and I would get inquiries that would say, I just want that site. Can you slap our logo on it? What’s the legal requirement?

for variation in order to kind of bypass the copyright laws. And I’m like, the fact that you as a lawyer are asking me this, I mean, come on, let’s just talk about red flags here. Let’s not do that way. Let’s back up for a minute.

Michael Buckbee (31:11.182)
Yeah. Well, so, yeah. So let’s concede that I think both of these strategies are not great. know, what, yeah. So whether you sit down and you hire someone else and they copy stuff and they add a little bit to it, or you have a machine, a computer do this, what we’re really recommending is to think about content in a slightly different way, which is that we…

Karin Conroy (31:20.012)
Yeah, especially as a law firm.

Karin Conroy (31:37.252)
Okay.

Michael Buckbee (31:39.212)
talk a lot about face and faceless content, which is that, so if you want to have good content, when you write it, it needs to be in someone’s voice. And like literally, if they read it and they can picture the face of the person who wrote it, then it’s successful. And if it’s not, it’s probably trending towards being answered by an AI search thing. And they can still both be valuable.

Karin Conroy (31:41.931)
what’s this?

Karin Conroy (32:04.654)
Yep. Okay.

Michael Buckbee (32:08.418)
But right now there’s not a lot of discipline about really thinking about like, well, is this face or is this faceless content? And I think there’s different approaches and different distribution methods for those, but yeah.

Karin Conroy (32:19.31)
Yeah. And as you said that, I’m picturing some of the people that I follow the most. And when they write some piece of content or put something out, it’s always like, I can hear their voices, whether I’ve got the phone on mute or not. And the really good stuff is clearly human. It’s clearly not just human, it’s theirs and they’re, you know, describing things or, and there’s, there’s nothing wrong in my opinion about

Michael Buckbee (32:38.925)
Hmm.

Karin Conroy (32:48.704)
I’m gonna write 70 % of this. Now I want AI to come in and overlay and kind of fill in. I guess that leads to my next question was, do you think it’s better to start the content yourself as a human and then have AI come in or the reverse or does it matter?

Michael Buckbee (33:10.888)
I think, I think AI is a tremendous tool for writing. Like I love writing. So, and I have written tens and hundreds of thousands of words, like probably this last month. Like, like I write, I write our own newsletter. I sit down by hand. I do curation of it, but I use the heck out of all these AI tools. And so partially that’s like, you know,

Karin Conroy (33:32.334)
Nice.

Michael Buckbee (33:35.638)
really sort of industrial like grammar and spell checking and content kind of stuff. But then also it’s a lot of, I think it’s great to have AI as a sparring partner, which is to say like, Hey, what am I missing from this? What else could I add? I think it’s also a great place to get started with a lot of things, which is like, give me the five things I should be thinking about this. And then, and I think this is, no, please.

Karin Conroy (33:54.106)
Yeah.

Karin Conroy (34:00.57)
So what’s your, sorry to interrupt, it sounds like your kind of very brief overview of your process would be to sort of inspire, like start with AI just to get some ideas and kind of thoughts around it, then human and then come back to AI. Is that kind of roughly how you do it?

Michael Buckbee (34:18.232)
Sure, so here’s a technique that I think works for a lot of different things, which is you ask the AI for too much. Like, hey, I want to write a headline for this. You don’t ask it for a headline. You ask it, give me 20 headlines. And then it will give you a bunch of variations and many of them will be bad. And a few will be good enough. say, you know what, give me 10 more like these. then you get, exactly, exactly. And.

Karin Conroy (34:31.555)
Yes.

Karin Conroy (34:41.338)
based on number three and four, yes, yep, yeah.

Michael Buckbee (34:45.814)
And that’s a very different technique than just like, just give me the thing, you know? And I think the people that are most skeptical are the, give me a thing and not the iterative process of this, where the iterative process is very fast, you know?

Karin Conroy (34:49.656)
Yes.

Yes. Yeah.

Right? It is. Yeah. We’re talking like less than a minute just to kind of what you just described. And I do that for this show all the time. I know I go in knowing like the general topic we’re going to be talking about. And the way that I come up with the show titles is exactly what you described. Give me 20 ideas on AI search and then it’ll come up with some. I’m like, exactly what you described. Most of them I’m like, no, no, no, no, There’ll be one that’s close.

Michael Buckbee (35:25.517)
Yeah.

Karin Conroy (35:27.628)
and then I go from there. there’s never anything that in the first result is what I end up going with. Yeah, okay. Sorry, that was kind of an awkward moment there. Totally. I do have the next thing. Did you wanna say anything else about sort of your process or anything there or?

Michael Buckbee (35:37.342)
So yeah, sorry. We’re waiting for the AI to chime in with 20 new topics to talk about. Yeah.

Michael Buckbee (35:51.214)
So if there’s one more thing about process, you know, we’re talking about writing, you know, I’m also a software developer and I think people who are outside of the world of software development are not, do not understand in their bones how much software development has changed in the last year due to AI assisted tooling.

Karin Conroy (35:55.011)
Yeah.

Karin Conroy (36:09.208)
This is fascinating. Okay, I want to spend a minute on this. And then there was something we were talking about before we started recording that I want to put in here too.

Michael Buckbee (36:17.836)
Well, so right now, if you write software, there’s a lot of dumb terms out there, honestly, like vibe coding, which is just like, just sort of say what you want and it does things. But the tools have gotten so powerful and are so useful that it is both taking the core work that I do and making it faster, like just making me more productive, but also in the same way that, you know,

Before everyone had a camera in their phone, they had in their pocket constantly, you know, the world took a lot fewer pictures because it was a more expensive, you had to make a decision and it’s like a dollar to take this picture and develop it and this whole, yes, insanity.

Karin Conroy (36:48.737)
Yes. Yeah.

Yes!

Karin Conroy (36:57.368)
Right. Have it developed. Yeah, buy the film. You only got 24 in your camera for the whole day. Yeah.

Michael Buckbee (37:04.598)
Yeah, you got to make them count. Yeah. And now I take 24 pictures of my food before you. So, you know, it so there’s there’s a massive change coming, which is that so many of these things that used to be expensive in terms of getting software written or having a special piece of software. My wife and I did this battle the books at the local public library and and we won. We beat like 15 other teams and we won because I cheated and used AI.

Karin Conroy (37:10.69)
Yeah, in four seconds.

Karin Conroy (37:27.555)
Yeah.

Karin Conroy (37:32.812)
Nice, nice, I love that.

Michael Buckbee (37:34.122)
I took all the books that we were gonna be talking about, just four books, and had AI break them up chapter by chapter into individual questions. And then I had AI write an application that was like a flashcard application for us. And it was easy. It wasn’t like a big deal to do that. And…

Karin Conroy (37:42.938)
I

Karin Conroy (37:47.793)
my gosh, I love this.

Yeah. Right. Yeah. It’s not like you sat and hand coded everything and like typed out character by character. AI just did it.

Michael Buckbee (37:58.491)
yeah.

Yeah, I just uploaded it to Claude and said, do this stuff and then spit it out. then, and it was a little technical and it took a little bit of time and it’s nothing I couldn’t have done if I had hired like a bunch of people off some website to read this and then break it in. But I never would have, you know? And, and so I think if there’s a real interesting part of all of this and where there is probably massive opportunity is in the places where

Karin Conroy (38:04.194)
Yeah. Yeah.

Karin Conroy (38:17.378)
Right, right.

Michael Buckbee (38:27.712)
Either there are legal things that used to be too expensive because they took way too much time, or there’s legal aspects where it just never made sense for a variety of these reasons of friction or you had to think about it too much. But yeah, a lot of changes.

Karin Conroy (38:41.146)
Yeah, I think this is the perfect. So the last thing I wanted to talk about is sort of looking towards the future, like a combination of like, okay, so what should firms be doing now and looking towards the future? And what I mentioned a minute ago that I wanted to just throw in here is this AI 2027. So I just got this from a friend yesterday. And if you go to ai-2027.com, there is this super fascinating

Michael Buckbee (39:00.75)
Yeah.

Karin Conroy (39:09.754)
report that was put out, I think in the last week or so, by the time this show goes, it may be a little outdated, but it briefly, goes through kind of where AI is now, and then a bunch of very well thought out reasoned predictions of through 2025, through 2026 and 27. And as you obviously get farther into the future, it’s, you know, it’s more hypothetical, but just looking from a global scale,

It starts a lot in sort of government and kind of international relations kind of thought. It is so fascinating. I just want to mention it because it’s some of these things were mind blowing to me and not things that were really on my radar. So let’s talk about what should firms be doing now and then let’s talk about sort of what you see on the horizon.

Michael Buckbee (40:03.726)
So I think two general things to keep in mind. One, things are changing very, very fast. And so many of the criticisms of AI and its use and its things are criticisms of AI from like a year ago. Two very specific things. The capabilities are much better on like a dollar level. If you can imagine some sort of abstract for every dollar you pay for AI help, what you get for that is a lot better. And

Karin Conroy (40:10.575)
Yeah.

Karin Conroy (40:18.851)
Yeah.

Michael Buckbee (40:31.598)
What you would have gotten a year ago for that dollar, now, I guess a better way of saying, what you would have gotten for $1,000 a year ago, you can now get for a dollar. Like the smarts have gotten that much faster, that much cheaper, that much quicker. So.

Karin Conroy (40:41.914)
That’s crazy. Yeah.

Karin Conroy (40:47.79)
Yeah, it’s just iterating. Like there was a time when they would talk about how fast it would take chips to iterate. And I think it was like a roughly 18 month kind of span on all that. And this is like such old data, but AI is like exponentially faster than, you know, what all of that, that was, you know, that whatever they were talking about years ago with chips and computers and all.

Michael Buckbee (40:54.883)
Mm.

Michael Buckbee (40:58.306)
Yeah.

Michael Buckbee (41:07.394)
Yeah. And there’s, is a very off topic now, but I think there’s again, so much of this is like an accident. And I think that’s hard to see from the outside because it is neat, but it’s still an accident. And like, we have all these computer chips that were never meant to run AI stuff that we’ve sort of tricked into running AI stuff and they don’t do it great, but we’re getting, we’re starting to get the chips that are meant to just like, all I do is run as you know. So.

Karin Conroy (41:26.82)
Yeah.

Yeah.

Michael Buckbee (41:36.598)
There’s going to be a lot of changes. so something I think about is like every form on your website, every single aspect of that data, there is going to be no reason not to run it through an AI because it will just make it that much better. And I imagine in the legal world, every single document is it’s just going to make sense to run it through some AI. Now it might be an AI system that’s in your office because there’s concerns about it getting out and it might not be doing stuff.

Karin Conroy (41:49.358)
Okay.

Michael Buckbee (42:06.53)
but it might say, hey, I know you’re gonna read this, but I happen to read it as the AI and here’s the 10 issues I found. Just throwing it out there if you wanna look at this too. And if it cost you a nickel, why wouldn’t you do that for every, you know? So.

Karin Conroy (42:13.11)
Yes. Yeah.

Karin Conroy (42:18.84)
Right. And so what would it, for an example, because the first thought I had was, well, we’re already all, especially lawyers, using AI like Grammarly, and that’s been around forever. But what are you anticipating that would look like? It’s kind of taking these documents or let’s say a case and throwing it up against sort of legal case history and comparing potential issues in the case or?

Or is it just more obscure than that? We don’t even know at this point what potential use cases there are.

Michael Buckbee (42:53.23)
So in AI search, there is a technique called grounding, which is that the LLM, the AI search is good at some stuff. But then it’s sort of paired up with a database. And so like you were saying, like, well, I found these 10 results. And then it sort of double checks them in Google’s index. And like, I can actually link to these. So I’m not hallucinating that there is a law office when there isn’t one in Seattle or something. So my.

Karin Conroy (43:11.032)
Okay.

Michael Buckbee (43:20.084)
It’s obvious there, and I’m sure there’s already some legal specific AIs that are backed up against Westlaw or backed up against other things because I know there’s been hallucinations and put before judges that have gone real bad for people. So not great. Talking, talking to the bar level bad. So, yeah. So I think that’s sort of the take on things is that, you know, we’re always going to have these sort of

Karin Conroy (43:26.764)
Right. Sure.

Yeah. Really bad. It’s like comical.

Yeah, right. Yeah.

Michael Buckbee (43:49.152)
assistive AI tools around us all the time. You mentioned Grammarly, Grammarly just shows up like you’re in a text box. You know, I think, you know, the ultimate is something the UX looks a little different. I think it’s going to look more like deep research where you give it a task and it sort of does a bunch of stuff. And five minutes later comes back with like, here’s what I found boss. And you’re like, okay.

Karin Conroy (44:10.584)
Yeah, yeah, awesome. All right, let’s talk about the book review section real briefly, because we’ve had so much to talk about with this AI search. But it’s time for the book review. If you are just tuning in for the first time, we have a Thought Leaders Library book section on the website where all of our guests recommend a book that you should have on your bookshelf. And so in relation to today’s topic, what is the book that you’re choosing to recommend to everyone today?

Michael Buckbee (44:16.398)
Sure.

Michael Buckbee (44:38.606)
So I’m a marketer, we’re talking about marketing and I think the basis of really good marketing is always understanding who you’re talking to. And the book I wanted to recommend was Deploy Empathy by my friend, Michelle. And it’s awesome. And it really walks you through not just like that you should do this, but the nitty gritty of like how you should conduct customer interviews and how you should do all these things. I…

Karin Conroy (44:47.672)
Yes.

Michael Buckbee (45:05.388)
I’m 100 % sure anyone who reads it will get something out of it. They’re like, yeah, well, I definitely paid for the price of the book. you know.

Karin Conroy (45:11.338)
Nice. I feel like the subtext sort of subtitle of this could just be like in relation to AI, like how to be the human, right? Like this is, okay, if I’m trying to go up against AI, what is the thing that I hold in my kind of toolbox? It’s that I’m a human and that someone wants to connect to me and work with me as a human. They don’t necessarily want to…

Michael Buckbee (45:21.292)
Yeah.

Karin Conroy (45:37.324)
do their entire marketing strategy through chat GPT. Well, maybe some people do, but like that’s scary. But the thing that I have that AI doesn’t is that I’m a human. like, if you go to that AI 2027 thing, there is some like cool stuff about AI trying to have feelings and, you know, do things that are more human like, but it’s still not human. So

Michael Buckbee (45:55.341)
Hmm.

Michael Buckbee (46:00.6)
Well, yes, I definitely double, triple, quadruple down on being human in the marketing. And for people that are, think, still on the fence about AI stuff, and I can’t imagine they haven’t stopped listening by now, if that’s the case. But if you’re still here listening, a great way to think about this is this is the new baseline. Grammarly comes out, the new baseline is you put out text without grammatical errors because it just takes care of it.

Karin Conroy (46:05.454)
Yeah!

Karin Conroy (46:15.159)
Exactly.

Karin Conroy (46:28.462)
Yeah. Yeah.

Michael Buckbee (46:30.146)
You know, the quality level of everything that we touch and generate is just jumped up because of all this AI assisted stuff. So if you’re not doing that, you’re definitely going to fall behind. But also if you can double down on your humanness and then add that to the tooling of the AI stuff, you can get much further ahead. And that’s my real hope as to what people do.

Karin Conroy (46:49.368)
Yes, I love that. Okay, so last thoughts in terms of looking forward and what do you see in the next year, 18 months that a law firm should kind of think about or have on their radar or be watching for?

Michael Buckbee (47:10.03)
I definitely think they should be looking for, the very least, can their sites be indexed by these AI bots? Does it have the sort of basic information that’s correct about them? And at the very least, start to think about how you can connect better with potential clients through a lot of different means. And in particular, I think that means producing human content and it means being real deliberate about that and getting your message out there as much as you can.

Karin Conroy (47:40.79)
Awesome. This was such a fun conversation. Michael Buckby is the co-founder of Noatoa, which really your main focus is this AI search and helping people understand and refine and clarify their AI search results. Is that right?

Michael Buckbee (47:53.134)
Hmm?

Michael Buckbee (47:58.604)
Yeah, so it will actually like you go and you start. You can try it for free. You go to the site, put it your website. And what we actually do is look up all the existing traditional search keywords you rank for. And then we break out, here’s the questions that had that same search intent for chats and for AIO and stuff. So.

Karin Conroy (48:08.544)
Nice.

Karin Conroy (48:15.34)
Amazing. Okay, that’s awesome. We will link to that on the show page. But thank you so much for being here. This was awesome.

Michael Buckbee (48:22.978)
Thank you. was fun. This is awesome. Thanks so much.Kn