Why Great Lawyers Get Recommended by AI

Episode 38Get Found by AIOctober 12, 202632 min

Why do great lawyers get recommended by AI?

AI search is changing how potential clients discover attorneys and professional service providers.

Guest: Adrian NikolovHaide Digital

Be prepared to Be-Picked.

Andrew Colsky, host of the Be-Picked Podcast
Hosted by Andrew Colsky, JD, LPCAttorney, licensed counselor and founder of Be-Picked.
Learn more about the host

Episode 38 · Listen now

Watch the episode

Brief summary of the show

AI search is changing how potential clients discover attorneys and professional service providers. In this episode, Andrew Colsky talks with Adrian Nikolov of Haide Digital about how SEO, GEO, AEO, entity optimization, and AI automation are changing the way businesses build online authority.

With nearly 18 years of experience in SEO and digital marketing, Adrian explains why traditional keyword-focused SEO is no longer enough and why consistency, credibility, topical authority, and strong brand signals matter in AI-driven search.

For attorneys and law firms, the takeaway is clear: being found by AI requires more than ranking for keywords. Your firm needs a consistent, credible, and connected online presence that clearly demonstrates who you are, what you do, and why you should be trusted.

See where your credibility stands

Take the free AQ Assessment. Twenty questions show how AI search sees you and what to fix first.

Take the free AQ Assessment

Key takeaways

What you'll learn

  • Why keyword-only SEO is losing ground to conversational AI search
  • How ChatGPT and other AI models evaluate authority and trust
  • What query fan-out and consensus mean for AI search visibility
  • What an entity map is and why it matters for law firms
  • Why your website remains the foundation of your online authority
  • How consistent business information strengthens trust signals
  • Why generic AI-generated content can hurt your visibility
  • How to use AI as a workflow tool without sacrificing quality
  • The difference between SEO, GEO, and AEO
  • How law firms can build stronger visibility across Google and AI search

Chapters

  • 00:00Introduction to Haide Digital
  • 06:31Authority and Trust in AI Models
  • 12:38Impact of AI on Paid Advertising
  • 24:19SEO vs. GEO
  • 29:41Reasoning Behind Content Strategy

Full transcript

Read the full transcript

Andrew Colsky: Welcome to Be Picked. This is the podcast where we talk with professional service providers about how to get themselves out found on the web so that they are the one that gets picked for the job. I'm your host, Andrew Colsky and today we have a guest, ⁓ Nikolov, and his company is Digital, ⁓ and is going to talk with us about a lot of the things you need to know related to SEO, GEO, and a whole lot more. So Adrian, to the show.

Adrian Nikolov: Thank you for having me, Andrew. It's pleasure to be here.

Andrew Colsky: Great, great. So why don't we start off with just the basics? Can you give us a summary of what type of things that you do in your company so that we better understand your background and things that you work with on a daily basis?

Adrian Nikolov: Of course, of course. ⁓ Well, I'm Adrian. We are based in Bulgaria and I've been doing SEO and digital marketing for the past almost 18 years now. ⁓ Hi, the digital was built with the idea to support organic growth for the future of search. ⁓ What we do and the way we call it, we call it organic growth engineering. And this is basically the combination between SEO, GEO and AI automation. We create systems that compound topical authority. We build semi-automated content delivery systems. And at the end of month five, each company can use the system by themselves. Absolutely no lock-in and a lot of AI.

Andrew Colsky: Okay, fantastic. Well AI is certainly a a topic of you know hot interest these days because it's showing up everywhere. So so let me let's let's back up to some of the basics. So you talk about SEO and GEO and then you talk about AI. So let's let's start with SEO. And maybe you can give us a rundown of what it is during your career that you have seen.

Adrian Nikolov: Yeah.

Andrew Colsky: SEO, how is it used? How was it used? How is it evolving? Give us sort of the big picture of where it was and where it is today.

Adrian Nikolov: course, SEO has evolved a lot over time. The way we used to search and the way we used to browse the search engine result pages was ⁓ you literally type in a keyword that you're searching for on Google and you choose which link to pick among 10 blue links. And that was it. SEO was thriving because it was way much, way much easier. There were a lot of shady tactics that got evaporated through several Google algorithms and the core update system. And nowadays it's completely different. So the way organic works now, it's not all about the clicks because search engines display a lot more information above the area where organic results surface. have AI overviews in Google. We have overviews in Bing. We have the so-called search everywhere optimization, which I will draw down on the topic later on in the show, but it's basically your brand needs to be present everywhere. You need omnipresence in order to get well cited and retrieved by large language models. And ⁓ speaking about large language models, the complete shift is exactly there. The user slowly

Andrew Colsky: Uh-huh.

Adrian Nikolov: But surely over these past three years, I can't believe it's been only like three years. They passed like three hours. But yeah, currently what's happening is that your potential buyers, your potential clients are researching mainly on large language models, comparing your company with others using large language models. And if you don't do the right job,

Andrew Colsky: Yeah, absolutely.

Adrian Nikolov: If you don't surface well enough in chat GPT, let's say, you won't get the traffic. It's simple as that. You don't exist. And the other thing, which is even more important, it is the sentiment, the way you actually appear in these large language models, because you know how chats with chat GPT go. You trust it. If chat GPT says so, so it is so. The guy is smart.

Andrew Colsky: Mm-hmm.

Adrian Nikolov: But what happens if chat GPT says that your brand is unprofessional, that the services you provide, ⁓ there's a lot of negative feedback and shady stuff around them. What happens then? This is even worse than being invisible. This is for sure that you're not getting everything and it's much harder to fix. Because authority nowadays, authority is, ⁓ I call it the new PageRank. PageRank is a Google's algorithm.

Andrew Colsky: Right.

Adrian Nikolov: Back in the days, was this battle. The page rank goes from 0 to 10, where 10 is highest. And we were all comparing, ⁓ how much page rank does this page have? a 10. Nice. I want a link from there. And they removed the algorithm. It's working on the backend. But currently, authority, the things that are said online about your brand.

Andrew Colsky: Mm-hmm. Yeah.

Adrian Nikolov: is the new page rank or at least I tend to think so and the data proves it. That's it in a nutshell.

Andrew Colsky: Mm-hmm. Right. So W that's that's fantastic. Excuse me, that's great information. ⁓ but can you tell us a little bit more? So I think you know, the concept of authority being the important thing here, especially for AI, you know, we're hearing that a lot ⁓ on this show. We hear that all the time. And so my question for you is when Chat GPT or any of the ⁓ AI models are out looking for you, what is it that they look for for authority? I know you said they look far and wide, but what type of things do they look for?

Adrian Nikolov: So, authority comes hand by hand along with trust. So there are a lot of trust factors. The way ChatGPT, let's take ChatGPT as a clear example, the way ChatGPT functions is ⁓ you type in your prompt and there's this algorithm, it's called Sonic. Their algorithm, it decides either to pick the answer from its core memory, from its training data, or to perform a web search. where this query fan out, the process is called query fan out, where your prompt is being turned into 10, 20 different prompts that are relevant for the model. ⁓ The large language model completes all of these searches, retrieves hundreds of pages, and surfaces only the ones that have adequate trust signals. What do I mean by adequate trust signals? It's positive sentiment around the web. ⁓ existing entity around the web, consistent information and consensus about this brand. Because I see it a lot of times with we work with several ⁓ lawyer companies and legal companies and they had this really ⁓ big issue that their information, the information about the company wasn't consistently surfaced around the web. At some... data points, there was information that ⁓ there were these types of lawyers that they handled these types of services. But on Crunchbase, their Crunchbase profile, said a completely different story. And when the large language model doesn't have enough consensus and reads all of these different information around the web, it doesn't know how exactly to surface you. So it doesn't at all. That's a major issue. So consistency along all your web profiles and presence.

Andrew Colsky: Mm-hmm. Mm-hmm. So, all right, let me drill down on that a little bit. So let's say I'm an attorney and I go out there and I list myself on all the different ⁓ places that I can find to list myself. And I say I do ⁓ auto accident law and I'm located in ⁓ Idaho and this is the law school I went to and I've been practicing for this many years and here's where I'm licensed. And I go and I put that up on all the directories that I can find and I go and I put it out on all of my social media. Is that gonna help me be found by AI or do I need a different approach?

Adrian Nikolov: Yes, that is going to help you. ⁓ Pretty much what you described is building your entity map, because you as a personal brand are also an entity. exist. having this consistency all around, being present in all of the trustworthy places such as Trustpilot, G2, and all of the other crunch base I already mentioned, and Clutch, ⁓ you prove to the models that ⁓ you are exactly this person. It's linked with images. You're using your images consistently and the model can connect you with these data. Thus it can service you. So it's more than enough. Then of course there is the caveat. If a lot of people with your name exists in the world, that's a completely different story. And there's the, that's the SEO battlegrounds out there, right there. ⁓ That's where you need to.

Andrew Colsky: Right.

Adrian Nikolov: up your game and start doing podcasts, start Wikipedia, apply for Wikipedia, which is the toughest part, but you can go and add yourself to all of the smaller wikis because large language models use exactly this for their training data.

Andrew Colsky: Mm-hmm. So is there is there a difference today because the average consumer has access to ChatGPT and other similar AI models? ⁓ The search that they do on AI, is that different today than the search that they would have done on Google prior to AI in terms of h search terms and what they're looking for?

Adrian Nikolov: Yes, it is completely different. The way search works right now and the conversation with large language models is conversational. You speak with it like it's your friend or your advisor to whom you're paying to. So we are speaking about much larger prompts. You're literally writing complete sentences, whether the way it used to be. On Google, we were... fighting for the keywords, for generic keywords. And ⁓ the intent behind each prompt and the chain prompting, the whole idea that you do a session of over 30 minutes speaking with GPD doing research. We haven't had that on traditional search. The way it used to go down was you just type in ⁓ the keyword, you click the first, second or third result or the three of them and compare them. And you end up there directly on the websites.

Andrew Colsky: Mm-hmm.

Adrian Nikolov: And right now, there's a lot more before that happens.

Andrew Colsky: Right. Huh. So is that in terms of the data, is that shift showing up? I would assume then that if people used to type in the keyword and then Google paid ads were based on keywords, if people are changing what they're searching for, then I would think the paid ads or at least that model based on keywords is no longer as relevant. Is there a shift that we're seeing?

Adrian Nikolov: Well, there is a shift, but not a major one. Ads still work. Advertising is huge. And that's once again, because of AI. The way ads function right now, it's once again, not only Google. Back then, Google PPC was the strongest channel. It still is for some more established brands, but it's getting expensive out of control. ⁓ Getting placed for...

Andrew Colsky: Uh-huh.

Adrian Nikolov: for a more generic keyword, what cost you around $50 per click. That's complete nonsense when you can invest in organic and be there for the long term, be there for the future of search. Advertising is going to come to AI. It already came to ChatGPT. We're seeing it. Currently, I'm doing an analysis on how ChatGPT ads are performing.

Andrew Colsky: Mm-hmm. ⁓

Adrian Nikolov: It's not looking bright, but I presume it's because the majority of people using it use paid profiles. ⁓ Ads are going to stay. For sure they're going to stay.

Andrew Colsky: Yeah, well they they show up everywhere eventually.

Adrian Nikolov: And they literally they hear us speaking. They know all of our interests. They're very targeted around ICPs and persona, which is great for a marketer. ⁓ I believe that's the proper way to do advertising.

Andrew Colsky: Mm-hmm. Mm-hmm. Wow. So okay. so I let's go back to I'm the this attorney and I'm trying to get myself set up so that I'm the one that's found out there. if in order for people to find me, if I just enlisted as an attorney doing auto accident law in Idaho, people would have to be looking for an attorney doing auto accident law in Idaho. But

Adrian Nikolov: Alright?

Andrew Colsky: If somebody has an auto accident and they go onto the web to search, they might be searching for ⁓ will my workers compensation pay me if I was injured in an auto accident? Or they might be searching for what do I have to talk to the insurance company about related to replacing my car from an auto accident. Th those are not pulling in the keywords that I would be listed under. So how do I how do I show up for those types of searches?

Adrian Nikolov: So this is the part where you definitely need your own website. By having this website, you can start building your own topical authority. ⁓ The term topical authority came from Koray Gubur from Turkey. ⁓ He is the grandfather of Semantic SEO. Topical authority comes from within your website, from the content you publish there. and it expands all over onto social media and everything that reflects back to you and so on and so on. The conversation outside of your website. The way topical authority is structured is usually by, ⁓ then, back in the days we used to do keyword researches. So you combine, you create keyword lists with all of the relevant keywords that this lawyer, this attorney should be present for. Right now we are more focused on topics which consist of such keywords but it's a bit different targeted. Currently we are trying to take ownership of the whole topic to become the subject matter expert of this topic. The way this is built is by having content structured around it, by having professional, original content and proof. ⁓ So first, you rank on Google. Secondly,

Andrew Colsky: Uh-huh.

Adrian Nikolov: you appear in large language models because you are the subject matter expert and you've gained trust. You're present in much more places. ⁓ In three, four months when you start sharing the articles that you write on your blog, people are going to start sharing them. You will appear in other media. This is another signal. And that expands and compounds. So... ⁓ I'm not trying to say that simply having a blog makes it better, but it's mandatory. And the other thing that's mandatory is separating these topics onto the intent behind them. Because YouTube, when we're speaking about blogs, we usually mean informational intent and people who are doing research, like how, what type of attorney, what I need if I end up in a car accident and stuff like that. But... The actual money pages that drive the conversion are the middle and the bottom of the funnel pages. So you need to work on those. You need to clearly define the services you provide. You need to add all of your certifications and the places we've just started with, all of the places that you already registered end up there as badges, as trust factors on your own website. And having your own website, if you're a solopreneur, Make your name your personal brand. Just do it. Connect the dots. Be omnipresent. Do a crunch base, everything that we discussed. And have this website. That's it in a nutshell. ⁓ Thankfully, we are in the era where this can be automated. And AI is our best friend here. You literally, you go to Clot or ChargeGPT.

Andrew Colsky: Right. Okay.

Adrian Nikolov: or whatever you're paying for nowadays, and you tell it, I am building this website. You give it context about yourself, about exactly what you do, even ⁓ very good practices that you make it interview you. You just type in interview me so we can get ahead and create a topical authority structure and the content that we need to build further down the road. I want to appear for these type of searches and so on and so on. You just communicate. ⁓ Easier said than done because ⁓ you need to have a bit of experience in SEO in order to surface the content properly because we don't want AI slop out there. People, even non-tech savvy people, are starting to recognize when a piece was ⁓ written by an AI nowadays.

Andrew Colsky: Uh-huh. Right. Uh-huh.

Adrian Nikolov: There are are a whole a whole lot of signs and it's kind of embarrassing. All right. I'm a I'm a professional I do everything I do this every day. I see it every day and I'm like, ⁓ no, not another AI generated comment. I'm gonna shoot myself and But yeah, this this things really matter so you need to establish your tone of voice and you need to provide it to the model as well and you need to test test test push the content out there and

Andrew Colsky: Ha ha ha ha.

Adrian Nikolov: study the data behind it because there are also algorithms. instance, Google announced, gave us a sneak peek behind how they are finding and how they are removing AI content from all around the web, like the mass generated AI slope pieces. And as Bert, the algorithm is working tremendously well. So if you just publish AI slope, you may end up... blacklisting your website, just completely removing it from Google. And if you're not in Google, that's a problem because a lot of the models use it, rely on it. For instance, ChatGPT, we've proved with data that it uses SERP API, which is directly hooked to the Google search engine result pages. So it pulls the information from there. ⁓ However, based on the Microsoft contract they had,

Andrew Colsky: Mm-hmm.

Adrian Nikolov: They also surface results from Bing. So Bing is another place that's growing very popular, by the way. They beat their records for traffic. Kudos Bing. Yeah. But I think you understood the picture. I hope I painted it right.

Andrew Colsky: Wow. Wow. Yeah, well yeah but can can you explain to us a little bit more when you talk about AI slop, how if if someone is trying to write a blog article and they use AI to help them, you know, with this, what differentiates AI slop from something that's not?

Adrian Nikolov: Yes. Well, AI slop may be described as an article which consists of irrelevant facts. ⁓ may add basically the AI slop articles that come out are ones that come out of laziness, the laziness of the person who is generating them. Because ⁓ you won't believe how often I see this. People just do go write me an article on this topic. It writes an article. They read the first paragraph, yeah, it's obviously good because it's written in like flawless English and they publish it either on social media or on their websites. But what they didn't go through was the fact that they are listing them and describing their service as a completely different thing. And even they may suggest competitors because we are talking about training data of the models and their

Andrew Colsky: Uh-huh.

Adrian Nikolov: pulling information all around from the web in order to please you. Large language models are people pleasers. So they're trying to get as much information from the web in order to please your request. And that's a problem as well. So you need to analyze and audit. The best way to prevent this is once again, I'll say it, provide enough context, ⁓ fix the tone of voice,

Andrew Colsky: Hmm.

Adrian Nikolov: and tell it what not to do. There are a lot of tactics how to do that. If you're using the desktop applications or the mobile applications, you can just add context files in charge.gpt. You can create projects ⁓ where you can add all of the specific context about this very project. ⁓ I myself use an ID ⁓ like like like developers do. ⁓ And I just

Andrew Colsky: Uh-huh.

Adrian Nikolov: I just create directories which are filled with these context files. That's like 0.1 % of everything that I do, but yeah, I just gave it as an example for the different types of structure. yeah, create projects, make sure you feed the proper context. You can also give examples with competitors. It doesn't hurt to say do better, but...

Andrew Colsky: Right. Right.

Adrian Nikolov: then you need to analyze how, what is the plan to do better and make it actually ⁓ deployable and efficient.

Andrew Colsky: Right. There's there's there's a yeah, I mean there's there's so much to learn there. let me let me make I want to make sure I cover this this other thing ⁓ too while we have a chance to talk about it. We've talked about SEO, but we hear now a lot different terms. We hear GEO, maybe we hear ⁓ AEO, different types of things. What is the difference between SEO and GEO, for example?

Adrian Nikolov: Never trust them. It is, is. course. ⁓ Different experts have various opinions, but the way I believe it is structured is everything is nested under this SEO pillar. Both generative engine optimization, which stands for GEO, and answer engine optimization, which is the exactly same thing as GEO. That's the battle. GEO or AEO, what are we going to call it? Just call it SEO. Of course, there are specifics.

Andrew Colsky: Right.

Adrian Nikolov: Each large language model has its specifics. It prefers different type of content. And Google prefers a different type of writing. For instance, there's this term that is called ⁓ BLUF, bottom line, upfront, that is popular in journalism. And guess what? Large language models were vastly ⁓ trained on journalist articles. So they are very familiar with this format. So get to the point from the very start. ⁓ What else? It slipped my mind. I just thought of something very important and interesting but it's gone.

Andrew Colsky: Mm-hmm. Well Right. Well, it it it might come back. I it's you know, as you're talking about that, it makes me think of you know, ⁓ there's certain websites, for example, that you go to and you're looking for a certain piece of information, an answer to a particular question. This this is this is not AI, this is just going to the website. And what people would do and they're all over the web, is they would tease you. They they'd write about it and write about it and it's gonna come and it's gonna come and Meanwhile ads are popping up and ads are popping up and you know, at the very end maybe they kinda get to what you were looking for and they force you to read through this whole thing so that you see all the ads and all the pop ups and personally I just go to a different website. I I I won't stand that stuff. But yeah.

Adrian Nikolov: Yeah. That's a completely different strategy. That's usually, I'm sorry, did I interrupt you? All right. That's a completely different strategy. that's mainly you see these on news article and on large magazines ⁓ or on recipe websites. ⁓ Yesterday I wanted to make a pizza. So how to do a margarita? I ended up...

Andrew Colsky: No, no, no, no, no. No, go ahead. Yeah. Right. ⁓ huh.

Adrian Nikolov: in on this article where I learned everything about Italy and the history of pizza, but I didn't just had the nerve to scroll to the the recipe and I just ordered. So yeah, this is a strategy that ⁓ is used for Google Discover traffic. Google Discover is the thing that is popular on Android phones. And the Google application, it's literally their newsfeed where

Andrew Colsky: Hmm. Yeah.

Adrian Nikolov: news articles and fresh content is being surfaced. It works completely different than traditional Google search and this is one ⁓ of the tactics there because scroll depth, ⁓ user engagement, like when the user is scrolling, it means that it's reading. He or she are reading the actual article which makes it valuable. They didn't bounce. Nobody reads why they are scrolling. the models still don't understand very well that they are scrolling because ⁓ they are looking for the actual information and then they suddenly rage quit. Yeah.

Andrew Colsky: Right, right. But I I just to me, now that they can go straight into a large language model and ask the question and get the answer, I would think that those types of pages might be in trouble. Am I am I am I right or is are they still doing well?

Adrian Nikolov: Yes and no. They are still doing well and surprisingly well. ⁓ Google discover and these types of pages are usually surrounded by a whole lot of clickbait. ⁓ They want to... They hit you right in the emotions and they really drive your click. You end up there searching because you need to find out if this ⁓ football transfer went well or who are we signing tomorrow or... who, attorney won this, ⁓ this case, stuff like that. And that's still working. And traffic traffic is flowing there. It's a bit less of course, but for this specific audience and these large magazines have their audience, the trusted audience, it's still working and it's still the, the, the largest piece of traffic they're getting.

Andrew Colsky: Wow. Wow. Wow. Amazing. There's there's a reason for everything out there. It's not just somebody being sloppy ⁓ in in developing what they're doing. They're doing it for a reason. So well we we have reached our time limit today and we've learned a lot of interesting information from you and I really appreciate you sharing that with us. If somebody wanted to get in touch with you, how would they do that? Do you have a website?

Adrian Nikolov: Yeah. Yes, it would have been fun if I didn't. You can always visit hide.digital. There we have a lot of interesting tools which me and the team have built that can help you analyze your own websites and how are they doing in either traditional or AI search. Check it out and of course there's a contact form.

Andrew Colsky: It definitely would be. Great. And so y can you spell that for us? Haide.

Adrian Nikolov: It's H-H-A-I-D-E dot digital.

Andrew Colsky: Perfect, perfect. Just want to make sure people are able to find it.

Adrian Nikolov: Yeah, and so if someone is interested, can always reach out to me directly on social media. I'm a very open guy. I accept anyone and I'm very friendly. So don't be afraid if you have any questions. Add me on LinkedIn. Adrian Nikolov.

Andrew Colsky: Great, great. Well, I want to thank you again for taking the time to join us, and I want to thank the listeners for listening in. Be sure to visit bepicked.com, be hyphen, p-i-c-k-ed-d dot com, where you'll find this show and all of our other shows, along with a lot of other information that can help you build and grow your practice. We will see you every week. We have a new episode. So thank you for joining beepick.com and we'll see you next week.

Left of Trust book cover
From the host's library

Left of Trust

Andrew's book on why excellent professionals get overlooked, and how to get chosen before the first conversation. Attorneys can read the first chapter free.

Read the first chapter

Ready to become the one clients pick?

New conversations every week on how professionals get recommended by AI search and chosen by real people.

Previous
Previous

How to Build Authority for AI Search: Adrian Nikolov on the Future of SEO, GEO, and Getting Picked

Next
Next

Google's AI Now Answers When People Search Your Business Name: What to Do About It This Week