Embracing AI for Success in the Future of Work and Search Marketing
[00:01:30] I think I'm good now. Okay. Today what I want to do is I wanna share with you what I see what I'm seeing as it relates to ai. Then I'm gonna share with you who I aspire to be from like a Seer standpoint. What do I want to
be now that I've seen these things? And then how can you as a client partner, Literally partner with us and help us to help you to win. [00:01:52] I've been saying this to my team, but it's like a game got reset and now's an opportunity for all of us. You, us included. And I wanna [00:02:00] be I wanna be there to help you all win, but it's gonna be, A partnership for sure, cuz the stuff has come at us really fast. So I believe this AI period that we're in, it's gonna basically change two things. [00:02:09] So these are my two areas of focus. My areas of focus are not on legality. My areas of focus are not on. Things like I don't wanna say internal buy-in, but I am focused on what is possible with these things. And then once we figure out what's possible, then we work
backwards. But if there's no value there, there's no need to me spending a ton of time with my legal team if I hadn't create anything in value in the first place. Today, we're gonna focus on creating value for clients specifically, and I think that the AI and the world that we're in right now is gonna change two things, how we work. [00:02:38] And then specifically how we find answers to things. Okay. So we wanna talk a little about how AI is gonna change, how we go about doing the work. Doing the work. And obviously folks,
I'm not an AI expert. None of us are right? But I do have a pretty interesting tenure. So I started doing search in 1999, August of 1999, and I [00:03:00] show you this image. [00:03:02] To tell you that before I even
started in search, people were already like SEO's dead. I got 23 years in the game. Seems to be doing okay. And then this is what search engine looked like when I started, there was relationships and there was like 20, 30, 40 of them, right? And I had to figure out what Open Directory project was and who did they power and who's as Chiefs, and who's Lycos and a o L, who's powered by a o L. [00:03:25] Is it Google? Is it Ink to me? What about AltaVista? What about direct hit? What about all these different sites? And I realized that I tend to perform best in areas of deep change in an industry, right? Look at that crazy chart. Of what we had to deal with back in 99 and 2000. And I feel like we're at a moment like that today. [00:03:44] And I feel like I'm at my best for myself, my clients, my team. When there is something like this disruptive happening
that has a chance to really be the future, but then over and over again as SEO grew, people kept pontificating on whether or not SEO is dead. Remember this guys, [00:04:00] oh, voice search. [00:04:00] 50% of all searches are gonna be voice. SEO is dead. How many of us use voice search for anything other than. Timers and music, right? So it's interesting when you work in an industry that people have consistently told you it's gonna die, and it really hasn't. It's changed, but it really hasn't. And God knows we're not gonna ever freaking do this in the metaverse, right? [00:04:24] We're not gonna sit here with these things on our faces, sweating.
Getting Zis so that we can kinda act like we're sitting together. So it's like there's all these predictions and we run around and we go, what should I do for the Metaverse? What should I do about the TikTok taking over the thing for Google? [00:04:38] What should I do about the voice search? And I like to really ground myself and like, where are the customers going? So of course everyone looked at search today and goes, do you think this is gonna kill SEO in search? And given how I've seen this job that I've been doing for 24 years, Being told for 26 of those years that it's gonna die. [00:04:53] This year it's starting to die. It's starting to die, and it really has not. So what do I think about ai? Is it just another thing that's gonna be completely full [00:05:00] of hype? I don't think so. I think this one is gonna be hyper,
hyper disruptive based on all the disruptions I've seen in search. I think that over the 23 years 24 years years, this is the most disruptive one I've ever seen. [00:05:12] And I think that it will actually impact the ROI of the investments that people make in seo. And I'm ready for it. Speaking of being ready for
things, the thing that's so important is where are your customers, right? Like we need people to find us and buy our stuff. So if you look at Web Van, they were launched in 1996. [00:05:34] They were a client of mine in 1999, and they were doing. Grocery delivery. I bet you many of us on this call today have had groceries delivered to our house, even pre pandemic, right? And it's great for a lot of us. So you think about it and you go the person that started, it started in 1996. Why aren't they around anymore? [00:05:52] They were too early. They started in 1996, IPOed in 1999, and were
dead as a business by 2003. [00:06:00] It wasn't because people didn't want to have their groceries delivered, it's because these companies invested way too early in the future of everybody's gonna want it delivered and people weren't ready yet. [00:06:10] Remember at this time, people were on dial up. Most people didn't have computers in their homes. We didn't have
that stuff. So therefore they had the idea, but they invested in it too early and failed. So to me, I think the goal that we have to look at is to follow our customers, not the hype. So the question is, and what I owe you all, and what SEER wants to lead in, is how to understand the inflection points when your customers start to change, how they seek you out. [00:06:38] Cause everybody's talking about this is gonna change, SEOs gonna die. It's but that's gonna be at different levels for different clients because
certain people's clients are gonna adopt it earlier and certain ones are gonna adopt it later. And we owe it to you to have the data at the ready for every one of our team members. [00:06:50] So we make strategic decisions and not decisions based on best practices and hype. So I'm playing around with this concept, right? But what if your monthly search volume [00:07:00] starts to drop for a set of keywords, but your rankings have stayed the same so you're not dropping in rankings. Your ana, when we look in your analytics, your number of leads from analytics stays the same, and your C R M is going, Hey, I'm still selling. [00:07:13] Then we go, okay. Maybe that word
people are starting to move from. Can I see if I'm getting more referrals from Bing through your analytics, are there ways that we can help our clients to know the inflection points at which their customers have changed their behavior? And can we know that fast enough so we can help you and we will know it fast enough cuz it will happen so that we can tell you exactly when that threshold is hit. [00:07:34] Imagine a world where you say, Hey, will, when more than X percent of my customers. Hit this set of metrics. That's when I want to know. Cause I need to start making different investments. Great. We can set up the alerting. We can find ways to make sure that you're notified when that happens. That's where I want to be and that's who I
want to be. [00:07:54] All right, so what data do we have access to today that helps us to understand how our customers are shifting? I'm gonna be [00:08:00] walking you through that, but before I do that, I'm gonna start using some acronyms and I don't want to To fire hose you on them. So I was, so I asked chat, g p T, how can I describe this to a five-year-old? [00:08:10] How can I describe it to a teacher? How can I describe these things to different job types? And I landed on one that is around a librarian. And again a lot of the slide deck was built with chat, chat, G p t as my thought partner. Okay, so when I get stuck, I go to chat G p t and ask it questions, and I say, how would you explain an l M to a lawyer, a doctor, a, this or that? [00:08:31] And I was looking for ones that were like, Ooh, that one I think people can gravitate to. So let's talk about librarians. I don't think we expect our
librarians to read and remember every page of every book in the library. Like we don't have that Expectation would be nice, but we don't have that expectation. [00:08:46] So I'm thinking of foundational elements. This should probably be one, two, and three. It should be backwards. But the foundation of what I'm gonna talk about today is that imagine a world where somebody read [00:09:00] every book in the library and they were able to interpret some of the meanings around those books and what things meant. [00:09:08] That's a good start. But then you need to be able to retrieve all that intelligence you just took in. And this is why Nvidia stock is up like 8000% or
something like that, because some money's gotta make it easy. Great, you read all the books, but when I ask you a question, I wanna be able to get that answer back in a reasonable amount of time. [00:09:24] And then imagine if you're a librarians says I read all those books. I can now retrieve all that information as fast as possible. So now I can write new books based on what I've learned. So when we talk about LLMs. LLMs or large language models are basically massive databases
of, all the words that in books, in Reddit threads and all these different places on Sears site on your site. [00:09:49] They crawled the web and got this information and found what words are likely to be around other words, the middleware on retrieval is the hardware. We are not gonna talk about that too much today. [00:10:00] But. What we need to be able to do is ac actually make sure that our teams can easily retrieve information.
[00:10:05] So the retrieval layer is what you hear about with all the hardware companies, a m d, Nvidia, whatever. They're making sure that you can get that information fast. But what's really cool is that once you understand the connections between these words, you can then remix it and that's generative ai. So now generate, for me, using your large language models, a rap song about Romeo and Juliet, where Romeo is method man, and Juliet is Mary j Blige. [00:10:28] And it does a pretty good job of that. Street to shall end. Yep.
Something method man would say. Even though rhyme cadence is very similar for method Man, and if you know anything about Bring the ruckus, bring the motherfucking ruckus, that song like they're talking about brought the ruckus to her balcony in the dead and night, you're like, damn, this is actually really good. [00:10:44] That's what generative AI is. It says, I have learned all of method man's lyrics. I've learned everything about Shakespeare, and I've learned everything about I've looked at all the lyrics from Mary J Bli and how she talks, and therefore I'm able to predict and write things and remix things in ways [00:11:00] that didn't exist before. [00:11:02] So you wanna read the books and understand the word connector. Probability. That's l m, right? Oh, method
man is highly likely to use the word shall in. Bismarck e not so much the tech layer we're not gonna talk about today. But that has to exist to make it accessible. So imagine a world where you read a million books, but then you have to carry them around with you. [00:11:21] You'd be like, no. If my goal is to recall what I've read, then my job is to get a Kindle so I can instantly search what I've read, right? So think of the tech layer as like the Kindle versus walking around with a million books, trying to remember I read them all, but how can I get it out fast enough? [00:11:34] Looking through all the books, looking through all the pages to find your highlights, and then making me new stuff based on the l. M is generative ai. And my friends size matters. Size really matters. So imagine if your library was these books and I said, write me this, method man, Mary j Blis, version of Romeo and Juliet.
[00:11:53] You'd be like, no, I don't. I don't have that. So I'm gonna do a really bad job of it. Okay. What if this was [00:12:00] your library? You're like, I might have a couple books. And then what if this was your library and you're like, oh, I got a lot of intel. So the important thing for you to understand is the size matters. [00:12:12] The bigger the library, the easier it is for something to validate that yes, these words are highly likely to be mentioned near each other over and over and over again. So then when you ask me to
remix something or write you something, I feel confident that I can do that. Alright, sEER has some of our own special books and I'm gonna talk a little bit about how important it is to bring unique data to these generative AI tools so that you don't get generic answers and sound like everybody else. [00:12:47] Cause if you think about it, these words, and you and all eight of your competitors are going into chat sheet PT and asking it to write a blog post about the same topic, all your shit's gonna sound the same. And we don't wanna be there. So now I wanna invite you in a little bit to our turf here, ATSR
[00:13:00] on something that we built years ago. [00:13:01] We've been working on it for years called Supernova, and it's our version of the library, right? Our version of books that your competitors don't have access to right now. Do I think they'll eventually catch up? Yep. But for right now, I would say for the next 6, 18, 24 months, you and I together as partners have a chance to put some damage on people who don't have access to this massive library of information we have. [00:13:24] Keep in mind, I know every word that you paid for, at least in your paid search campaigns, used by your customers every day for the last two years. I have it all. So I can see trends in there.
I can see how your customers have changed their language. I can be like, yo, chat chip pt, given that my customers are starting to use these words more, how would you write this? [00:13:42] Right then Supernova not only takes that pay data, it joins it to the organic data instantly and allows us to then go, Hey, you know what, Google, if I gave you the best, the number one ranking page for this set of keywords, could you then use that as the model [00:14:00] and chat sheet PT to write something for me? [00:14:04] So I'm combining how my customers chain language I'm combining at in real time. What's the number one result? Then I can take other things like the people also asks and other things, and I can use all that just sits in our library at a very massive scale. And then give it back out to chat g p t to try to come up with different and more unique answers that are how your customers are talking. [00:14:26] And also the questions that Google's predicting they're highly likely to be asking based off of Google's AI and machine learning, which they've been using to power. People also asks for seven,
eight years. And then imagine I got all this stuff in here. I got competitor ads, I got titles, descriptions, meta descriptions. [00:14:41] I got all this stuff sitting here. So then as a partner, I want you to bring us challenges. Hey, if you've got all that information from the search result and I can't get it through most of these other tools that I have, could I use that to generate answers this way? Those are the kind of things I want you to bring into my team. [00:14:58] So Supernova is the base. [00:15:00]
We use Looker as our interface to that base of data that we now have. And then our consultants are gonna be the librarian who can now remix. It's a combination of technology and people that let us run and to create an opportunity for our clients, I think for the next six to 24 months, depending on how fast the industry moves. [00:15:20] So what I think is table stakes
is this stuff. If you're saying, Hey Bill, I wanna go into chat CT and have it write me about 20 cooking techniques with hot oil. I'm not here for that. This is gonna be table stakes. This is an expectation. And I will at some point we can talk about like the table stakes basics, how'd you write content, et cetera.
[00:15:36] But if we're just gonna write content without putting in our own unique information from our library, then what are we actually doing? We're just putting out low quality stuff that Google's gonna eventually kick out a bar or Chachi PT will, or whatever. The best marketers are gonna be saying, how can I prove the answer from chat c p T, or Bard or Perplexity by feeding it my personal information, not my personal information, but the [00:16:00] information that only I have that my competitors do not. [00:16:04] But it always starts with the customer. This is not about technology. It's not about hype. It's about how do I understand how your customer, what your customer's looking for, and then how do I apply the AI to better speak their language to them, to help them convert better, quicker, faster, and cheaper for you to either drive up your sales or drive down your cost to get those sales. [00:16:23] So this is a search result when I talk about Sears Library, okay, I'm talking about all this stuff. We know if an image shows up in the number one
result, And where it shows up is important. Oh, Google knows that HIPAA number one result should be an image of moths. It's a type of moth. Then we've got the other things that people searched for in our library. [00:16:42] Then we've got what sites are showing up in our library. Hey, can you write me a page about HIPAA that's very different than W Wikipedias page. I can do that, right? But then you got all that and it helps us to use just
basic data and AI to make sure that [00:17:00] our clients don't end up doing stuff like this for the ones that we are managing. [00:17:04] Your paid campaigns, right? If I got all these clues in my library that you can't find in Google, let's go back. Google will not tell you that we know that this an image of a moth is the answer for hipaa. They will not tell you that, right? Because they're not willing to tell you that anywhere in their tools. [00:17:22] You can show up for things like
this in your paid results and get clicks you didn't want, and it's gonna drive up your customer acquisition costs. We have this pretty much an as an automated thing that we can run across all our clients. And if you're not getting those insights, let me know directly because we've invested millions in the technology to be able to make sure that you can get that on a week, on a monthly basis.
[00:17:42] I actually check this for all of you most of you on a monthly basis and send out internal notes to the team on things that I find. I've been doing that now for a couple months. So imagine a world where we have to build new landing pages and everybody's gonna have a way to use ai. Oh, build me a landing page for this. [00:17:57] Sears Library. Let's just say it
has [00:18:00] 700,000 ad copies going back two years, and we've got all the pages that those competitors of yours landed people on. That's in my library instant. You can go into SEMrush and try to pull some of that down or go to other tools and try to pull some of that down. [00:18:14] But we find that our data set is extremely, when I say much more robust, I'm talking about like a 10 x to a hundred x more search terms in paid ads, in paid landing pages, in paid, et cetera. It's a massive difference. And remember, size matters. So if somebody that you're gonna use has very few
resources to, then you're gonna put that in a chat G P T. [00:18:34] But I've got hundreds of thousands of ads. Sitting in my library. Then our version of using chat g p t is gonna give better answers that are more accurate than someone else's. Alright? So I just was playing around and said, all right, I gotta build a page about PPC agency pricing. All right?
[00:18:51] Remember in my library I've got the ad domain, all the ad copy, all the site links. Imagine if you said, Hey will what are all the [00:19:00] site links for all my competitors that include this testimonials? I can do that, right? We have this information sitting here for a good chunk of the ads, but then I've also got the related searches from seo. [00:19:10] So then I could say what are people looking for down here? Estimators, hourly rate, what is it? So I can use the related searches from my library to feed it to chat, g p t to go. These
are also the related searches. How would you write a better ad for me? If I do not have that related search box parsed out and at the ready for my consultants, they don't have that in their library, which means their answers may not be as robust. [00:19:34] And where do you think these related searches come from? Google's wickedly smart machine learning. They update
their results in real time folks, which means Google's gotten really good at understanding what people want and how to update results quickly. So we want to use all that intel and harvest it and put it in our library to help you all to be successful so that when we go to chat, g p t, imagine a world where I go, Hey I'm trying to build a page about PPC agency pricing. [00:19:59] I'm gonna [00:20:00] give you three ads and three landing pages. Use the link reader. Plug in to index the pages and tell me if the page is a good response to the query. Do you understand? Do you understand at the end of your request is a very interesting thing to ask chat pt, cuz it'll show you where your queries are not clear. [00:20:17] So you can learn how to be a better
prompter of chat sheet pt. So I throw all those ads in, it goes, yeah, none of them. Imagine if we could do this at scale for you. Take all my competitors' ads, take my top 10,000 words that I spent more than this much on. And tell me how well their ads actually answer the query and I, and to do that in real time, deliver me a monthly report, whatever.
[00:20:41] It's, that's what I wanna work towards. But then the kicker is, see how I've got three links down there. There was a plugin for Chat g, PT four that then let me put in all their landing pages and it crawled all their landing pages and gave me a summary to tell me whether or not those pages were actually about what the query [00:21:00] was. [00:21:00] Because the query is what the customer wants. The clues and organic search is Google's best prediction at what the customer wants, and that is why when you search for HIPAA with a why the organic searches are like, this is for a moth. Google knows what people probably want better than most, any other thing that I know of.
[00:21:18] So then I was like, Ooh, this is getting good. Now let me take the number one ranking page in Google for this keyword and go crawl that and tell me what you think. And it goes, whoa, this did an excellent job. And let me tell you all the reasons why it did an excellent job. Imagine if we did this for every one of your clusters of search terms where you spent over a certain amount of money. [00:21:36] Or anywhere where your C P A got
really high for certain types of words. Imagine what we could do if we did this, but we're not done yet. I then was like, eh, you know you're a client. You probably are like, I don't want these paragraphs, man, can you put this in a table? I just asked to put it in a table and it did it. [00:21:52] The ratings aren't so good. But then I went, okay, so now I'm getting ready to write my page for Seer. Do I just start writing it? [00:22:00] No. You
put in your customer objections. Imagine if you powered me and my team with your chat logs or your, or when customers quit or when customers have a bad experience or when they have a great experience. [00:22:11] Imagine if you said, that's why we gotta be partners in this. Cause I got my library ready to go. I wanna help you to get your library ready to go so that we can make library babies together, right? So it's like I take my library and you take your library and you find a way to wrangle the team internally to get us chat logs, to get us reasons that people aren't picking you or whatever. [00:22:31] And now we can put in all of the objections into chat, G P T. That's why you gotta bring unique information to
it so that you get better answers. So in this instance, I said, Hey, let's just say we're expensive, but here's some of the pros and cons. And then I said, because I now have a plugin that can crawl the web. [00:22:47] If you need to get my tone on how I talk, check this page out and then write your ads. So it went to the page, and then it wrote me the ads. Now I can use [00:23:00] my team of consultants to work on the tweaking and the understanding of your business, not on trying to type the ads in. All right. Yeah, this is good. [00:23:10] I'm really excited about this opportunity
that we get to take advantage of together for you and on your behalf. So I'm gonna focus mostly on customer friction, on like, how's it gonna change how we find stuff? So let's walk you through finding a car under 25 grand. And this is just one website. The price of new cars has risen substantially. [00:23:29] All right, cool. That's not what I asked. I didn't ask, did the price of new cars rise? That's not what I asked for. And I didn't ask for an image of a
civic, but you're gonna give it to me all before I can even start to get the real answer. And then I come down here, but don't get discouraged yet. [00:23:43] There's still lots of options. What? That's not what I want. And then you're gonna make me click 25 different times to go through the top 25. So now
I'm on the Subaru Legacy page after 20 clicks, and then I'm thinking, did the civic have standard all-wheel drive? Now [00:24:00] you're gonna make me click back 19, 20 something times. [00:24:03] It's like, how am I supposed to remember all this stuff from all the things I clicked? We're not LLMs. We have very limited memories and we have a very hard time retrieving what we have digested. So now we're adding friction for our customers, and this is what I might really want to know. [00:24:20] And therein lies the power of Chachi pt and why I think it's insanely disruptive, because it's gonna be disruptive on how we find things because we have too much crap around the answer. And that's why I think customers will probably move and adopt. So I wanna know about
safety. I got two kids. I wanna know about safety. [00:24:36] It gives me the answer, great. But then I'm like, eh, I live in Philly. Sometimes it could be a little bit cold in the mornings. I need heated seats, and I'm a Samsung guy, doesn't have Android auto boom. Gives me the answer.
Then I'm like, eh, put it in a matrix for me so I can compare all these cars to each other. [00:24:49] And here's seven other things that I think are really important. Great, thanks for that. No ads, no popups, right? And then I go, Ooh, those higher trims. That's not helpful. [00:25:00] Tell me what the exact trends are. It doesn't great copy paste, put it in my email. Now I know what to get, but I'm not done yet. People do different things with cars. [00:25:10] So now I can say, I live in Montana and I mountain bike in the summer and ski in the winter. Which of these cars would be best to keep my
equipment in for these activities? Oh, then you want the Impreza, it's got this much cubic feet of cargo space and they're like, broadly skis are this long. Broadly, I have understood how big a typical mountain bike is. [00:25:27] L m. I understand how big a typical mountain bike is and I
understand how big typical skis are, and I understand that when you say I ski in the winter, you don't have to say, I have skis that are this big. So it's understanding all of that and now it's saying, oh, based on the cubic feet of all of the different cars, this is the one that has the most space for you to do the things that you specifically do. [00:25:51] This is why I think this is disruptive. And I'm gonna show you in a bit how we are trying to lead in this regard, but I'm [00:26:00] not there yet. I want to remove friction for customers. You as my client, I wanna remove friction for you, but then I want to help you remove friction for your customers. I wanna find the places where we're putting people in the wrong place and then use AI and machine learning and chat, G P T, and all these different tools to help give your customers better experiences.
[00:26:18] Cause I think it's gonna lower your cost and help you to convert more customers. Imagine a world where we're trying to pick a Medicare plan, we're gonna watch New York living, and they're gonna tell you about it. That's one way, but we're all personal and have our own things that we might be looking for. [00:26:32] I don't want a TV ad. I want to say, Hey, what are the copays for all the plans? Then I wanna say, Hey, here's my favorite doctors. Which ones are in network for the different ones? Give it to
me in a table. Great. And then I'm gonna say, can you gimme all their ratings, copay, and distance from my house in Erie, Pennsylvania? [00:26:50] And it goes, sure, I'll do all that for you. [00:26:53] Did you catch the hallucination? Let's talk about hallucinations a little bit. These are things are gonna happen, but I'm gonna tell [00:27:00] you why I am not. I'm concerned. But that's not where my focus is, and I'll tell you why in a second. But do you see the patterns? This is what I'm trying to teach my team, because this is where we need to know where Chepe T and Bard suck, so we know exactly where to use humans to do the work.
[00:27:17] Because otherwise it's gonna be wrong for your customer, right? That's what I'm trying to train people to look out for. Notice how it goes. 4.5, 4.0, 3.5. So it's just taking 0.5 off of each one. Pattern matching. All right, these guys are up to something. And then the rating is a dollar sign. That's wrong. [00:27:33] But then it's 10 plus five, 15 plus five is 20. Copay, 15 plus five, 20
plus 25, 25 distance from my house, 10 miles, 15 miles, 20 miles. You're like, oh, this is just making it up. And if you're like that's why I don't wanna use this stuff and you're gonna miss out. Bring us your hallucinations. Email me. I want them because then I can learn how to better train my team to find them.
[00:27:56] This is gonna be a team effort, y'all. It's you and us together, the more [00:28:00] people trying things, working on things and whatnot. I'm sharing with you everything that I got going on. To hopefully inspire you to start doing some of this stuff. And then in turn, I'm hoping that you will be like,
yo, I played with this and this really didn't do it for me. [00:28:11] Or, Hey, see your team. What have you learned about this or that? That's what I'm hoping for. Now, if you look at those hallucinations and go,
that's why this isn't gonna work. Let me show you what happened with Mid Journeys prompt for Selfie of Yoda. If you saw Mid Journey version one and said, this thing sucks, I'm never using it again, then you would not have been there one year later when it nailed it. [00:28:32] And that's how you're gonna be
behind. So please do not use hallucinations. If anybody wants to talk to you about hallucinations and how bad they are, get out of the room because those people saying that also have no idea at the processing power and the ability for these things to learn. And you don't wanna be on the wrong side of it getting better. [00:28:51] And of course, I asked Chad g p t, I'm like, how would I get a Yoda quote? And it was like, oh, here it is. It's
actually funny. At first it made up Yoda quotes and I was like, and then I Googled it. [00:29:00] And I'm like, none of those are Yoda quotes. Can you give me actual Yoda quotes? And they gave me this one. [00:29:03] So be patient. All right, so now I wanna talk about some different
roles and how we are specifically incorporating this stuff and how we're thinking about incorporating this stuff. Some of this stuff is super alpha, so this is where your feedback on this webinar is critical. I cut out on 50% of my slides because I got so much more content to share with you. [00:29:21] And I need to know, do we have to do another one in a month and another one to keep you guys up to speed or are you getting everything that you need? Some of this stuff is alpha, some of this stuff we can do right now today.
I also believe I owe it to you to not say we got ai. I need to explain to you how we are driving unique value to you and how we are uniquely doing something that I don't believe I have seen anyone else doing. [00:29:43] That is how we create wedge opportunities between you and your competitors, and that's what I wake up to do every day. Drive a wedge between you and your competition. So now I gotta figure out how to use these tools to
make that a reality. If I search for the word, Brick fireplace or brick ideas for brick [00:30:00] fireplace and Google shows a bunch of images that show fireplaces and bricks in a can. [00:30:05] That could be a bad answer for the customer. So I've been at using visual AI for almost two and a half years or so, right? So I went, wait a second. In my library I currently have whether or not an image shows up, but everybody else has that too. I'm not here to be like everybody else or build a company like everybody else.
[00:30:24] So if Conductor has it, if SEMrush has it, then what am I doing? Building something that they already have, right? But then I went, you know what? They don't have what's in those images that's similar. So now we have built a script. If you're interested in this, you need to mention it in your comments. [00:30:44] And by the way, if you're also
gonna say, Hey, will, what I would prefer is for you to tell me how much did I spend last month on keywords where an image was in the top three. I already have that built now, and I can deploy that to my team, and they can send you those reports within a matter of minutes [00:31:00] because we've got the, we've got the retrieval mechanism figured out to make that an a legitimate reality. [00:31:06] Okay. So imagine if I take all the properties of all your search terms where an image shows up in the top, and I can say, let me bring in your images from your landing pages and see how close they are to the images that Google is actually rewarding. How far off. Are the factors around your images versus the
one that Google is rewarding. [00:31:27] And what that allows us to do is go oh, we can see different things, rectangles, woods, crass, whatever. So then that helps us to figure out, do we want lifestyle images? Do we want images with little squares? And it tells it to me by search term. So now I can go for these search terms. You need some squares for these search terms. [00:31:42] You need lifestyle images. Why are we showing squares to people
who want lifestyle and vice versa, based on what Google's already learned, that I can suck out of their ai. Now, this kind of stuff. It costs us money to run these and whatnot. So we have to figure out if you want it and you wanna be on the cutting edge with us. [00:31:56] There's real cost to use these
tools, the vision AI tools to run this [00:32:00] on hundreds of thousands of words. But I will tell you, I don't know anybody on the planet that's doing this, and therefore it creates a wedge opportunity between you and your competition. And that's sitting in my library today. [00:32:10] Alright. Let's look at content. So everybody's oh, can I use chat two PT to write content? I'm like, that's the freaking table stakes. Instead, what I wanna do for you is things like this primary care physician, so I want all your competitors to go. I need a page about primary care physicians in my city.
[00:32:28] Cool. I want all of them to do that. I wanna use the data I have today to tell you, hey, the word black in front of primary care physician has increased by 300% over the last two years. People, your customers have changed how they're seeking out primary care physicians. That's what I want to do for you,
and how much did I spend? [00:32:47] Now you've been spending $55,000 a month on words that include black in front of different doctor types. You go, I would build a new landing page. If I'm spending five, five grand a month, say 60 grand a year, I would build a new [00:33:00] landing page. Great. Then I can say, show me all the words around primary care physician, where you have spent over the amount that you told me you would build a new landing page that is available to day in supernova right now for me to do or for your team members to do. [00:33:14] You gotta send in the requests.
We can get you the answers. Cuz it looks like this. Oh, when this word was used, all you went from a CPA A of 1 38, 12 months ago to 2 44. Now maybe we are not answering that question well, and that's why it's costing us more to convert those customers that's in the tool today.
[00:33:34] But then let's think about what we can do if we do the other parts of the tool that are there. So it's not enough for me to say, Hey, this word showed up. What I also want to do is take your URLs and go, here's your top URLs. Here's the top URLs that Google is ranking for. Those same, for that same basket of 119 keywords. [00:33:54] Those are the people I should probably copy, right? So let's go back to a black primary care physician [00:34:00] search term. Imagine you are a marketer at a healthcare plan, and you, I just dropped that information on you, and you're going, that's a I should build a page. What does the right page look like? And then I show you your competition
for that keyword. [00:34:13] Anything that's related to black doctors here, this is not a game. AI is not a game about who's got the most money. And I like to work with underdogs. So this is about who's gonna go get it. Zoc Doc's got more money. Than my clients do in this space, but they didn't got the tools that we got, which lets us eclipse them and find areas that they're asleep at the wheel. [00:34:34] That's what I wanna do for every
one of you guys. So I'm gonna pull in Zoc Doc's page search ad, which I have at some of their ads, not all their ad and their landing page into my tool set. Already doing that today. So I can take this page and go, that's not great. I can take this page and go, there's part of a black doctor at the bottom.
[00:34:54] But if my search is for black doctors, why are you paying to drive your customer to a page that's got white doctors on it? And then you've [00:35:00] got the Amazon page with one medical and you might go, oh, they've got a black person there. This is not about, that is just inclusivity imagery. The page is not about black, the black experience going to a primary care physician. [00:35:16] So what would we do? Partnered together? How can I help you to beat these three competitors? I just showed? I've already showed you that. I got the tools that will tell me every month all the different words around the word primary care physician. I can trend out your C P A and tell you how much you're spending.
[00:35:29] I already have the ability today to let you get an alert whenever that number gets high enough that you're willing to work with my team or your team to create a new landing page. All that is already done. Now, let's throw AI on that and show you how we can supercharge that. Because we don't want that friction for our customers. [00:35:44] So I go to chat, g p T and go. If I wanna better communicate with people who are searching Google for Black doctors, what kind of messaging should I use? Highlight your diverse staff. None of the competitors did that focus on the importance of representation.
None of the competitors did that and show an understanding of health disparities between [00:36:00] black communities and other communities. [00:36:02] Nobody did that. I have the ability to crawl their landing pages, find out that they're missing it instantly. Then run and have a team that uses these tools to go, okay, we found the word is off. We found that none of these people are answering it right? We're gonna ask chat, pt, how to build a better page. [00:36:16] Hey, client, let's build that page. I'm gonna hopefully drop your CPA
in half. Lower your customer acquisition costs, and speak the language of the customer. Drop the friction and let your competitors keep getting all these clicks that aren't answering the customer's question, which means they're just gonna be led to us anyway. [00:36:29] Maybe we could even build a little
bit lower bit, a little bit lower, and then we can do this at scale and be alerted whenever these things trigger at a level that you're willing to take action. No more recommendations where you're like, that's a good recommendation, but it's not spending enough for me to care. [00:36:43] Tell us how much it needs to spend for you to care. And we can make sure that we only tell you the trend when it hits the thresholds that you care about. That's in the, that's in the product we've built over the last two years today, and hopefully you can see how that. Unique library helps us to beat competitors, sleep at the wheel using ai.[00:37:00]
[00:37:01] It's funny, everybody talks about ai, but AI is only as good as the data it was trained on. The bigger the library, the better we can be. So Bing has started to look at, yeah, we're doing good on time. Cause I'm gonna leave the last 10 minutes or so for questions. So Bing is looking at all right. We were the
first people out with really a chat search outside of perplexity and some other folks. [00:37:19] But they were the big splash and they're now starting to disclose where they're seeing people happier with chat and versus where people are happier with a search experience and you can see best coffee machine, best how to, I need to throw a dinner party. Can you suggest right, that's where chat wins. [00:37:41] Things that are more specific, search typically wins. So this is gonna be really important for you to lock up your
branded search because that's what people are gonna go searching for. It's really important for you to lock up your branded PaaS because that's gonna show up in the future world. And this is group being's early data. [00:37:58] We have the ability to kick out for you, that [00:38:00] kind of thing at scale. If you're not already getting it, let me know cause we have already built
it, so therefore we should make sure that you're getting how, what percentage of my branded people also ask questions, am I not owning? Because in the chat driven world, that's gonna hurt you a lot more if you're not one of the answers for things related to your own brand. [00:38:18] Alright, so I'm gonna, this is a prototype. This is a prototype we need your feedback to know if this creates value for you and if you would do something different than you're doing today with this kind of information. I believe you will, but I need that feedback so I know how to prioritize my engineering team's time. [00:38:33] So it's really important. I want to be a partner with you in this and I want you to give me the feedback so I know where to put my amazing team on problems that are gonna give you the opportunity to drive a wedge between you and your competition. We can't do it all at the same time, so it's really important
for your, for you to come in with feedback. [00:38:48] So remember when I mentioned earlier that we built a way in Google Sheets and shared it with the whole world, put it on the blog. In a world where more people are gonna be searching for your [00:39:00] brand I'm like, okay, and in a world where rankings go away rather, so let's just say people are inside a chat. [00:39:05] If today you're thinking, okay, I know where my ads are, I know where my rankings are, and all of a sudden if 50% of my people are using chat, what are what's a ranking in that world? And am I visible? And I was like, huh, I would wanna know that because people are still looking for something. They're just using a new channel to do it. [00:39:23] I built a little template and allowed you to put in a couple questions and it was good. But then my partner Jordan
over here, he's sitting next to me over here he was like, I'm gonna take that and blow that out. I'm gonna take 10,000 questions and I'm gonna use chat, G p t and some other tools to help me to blow this out for how we can help our clients to know where they are and aren't visible. [00:39:42] And I'm gonna go ahead and hand that over to him right now. So Jordan I'll drive. You can go ahead and take it away. Awesome. Thanks Will. Yeah, I'm super excited to show this off the thinking about what Will's been going through. We have the library of data here, right? And as part of that library, [00:40:00] we've got questions that Google thinks that customers are searching for, and we have cost metrics on what you folks are spending to show up for those different questions, right? [00:40:12] So in this example, We searched for we're looking at CRM clients, right? So what questions contain crm? Now we filtered to that. Now we can see, all right, of those questions, which of them
cost the most last month, right? So we got this big list of questions, 135 K spent, and then we took the chat, G B T data, blended it in here for these questions, and now we can see. [00:40:39] Who are the top brands that are showing up for these CRM questions? And we can see Zoho, HubSpot, Salesforce, there's some of the top brands showing up for those. And if we want, we can take a look at the exact responses to get some context on why they're showing up next to each other and what GPT is doing it.[00:41:00] [00:41:00] And what I wanna highlight is what's going on here is bringing together those three data sources, organic, paid, and chat G p t really gives us an edge to say, Hey, where should we be focusing on that's gonna drive the most value for your business? And let's get some answers to those things and see who those different competitors are. [00:41:20] All right. So we figured that first part out. The second piece that we can do is in these different. That are coming up in chat gpt in a world where that's becoming more frequent, let's say that I'm HubSpot. I'm a CRM competitor, and I wanna know when I show up
in an answer, who else is showing up next to me? [00:41:42] And here you can see in the example of Got Hub site searched for Hootsuite's showing up in those answers. And Salesforce is showing up in those answers. So if I'm thinking about who are the competitors that I might need to be concerned about or taking a look at what they're doing to show up in Google search in these answers, this is gonna [00:42:00] help show me that information. [00:42:02] And then finally There's really a chance for disruption here, and this is an example of where something like that's happening for these CRM questions. When you look at the responses in Chat G P T, you can see that pipe
drive is a CRM that's showing up 36% of the time, whereas the big gorilla in the room, Microsoft they have the dynamic crm, they're showing up 31% of the time that you'd expect for. [00:42:31] A big competitor like Microsoft to be showing up a ton in these, but Pipedrive the small guy, they're showing up more frequently in the chat TPT responses. And so what I'm really excited about as this moves forward is we'll be able to see this kind of information over time as we track it to say, Hey, as generative search becomes more of a thing, as that starts showing up in the Google search results, who's showing up more? [00:42:55] Are we still owning those results? Is there a chance for some of the smaller guys to [00:43:00] gain visibility there as that becomes a bigger part of the search landscape? [00:43:04] There we go. All right. I was trying to draw on your slides and keep people following. I know we're throwing a lot of data at you all. So I'm trying to do my best to to make sure we don't overwhelm
you. So if I could recap, what we did is we looked at what Bing said people are starting to look for, and they're like those best queries, those how-to queries. [00:43:19] Those kind of queries are the ones where they're like, chat is a really good answer. So then we said, let's take all of those from the millions of
questions that we have in our library. And then say, let's just cut a 10,000 cutout and then say, where does, how often does our clients show up? So when chat sheet PT is making recommendations about what companies are the right companies for you to think about, which will drive branded search later, which is where they will end up on search. [00:43:42] We want to know if our client's not being mentioned in those places, and how much are they spending on those kinds of questions. All right. So there's that. And I wanna tell you that we are in a hyper disruptive time and obviously I've gotta get an eagle shot in here. I want you to never question the effort that the [00:44:00] team is going to put behind, trying to make sure that you take full advantage of the opportunity that's ahead of you ahead of all of us because it is an opportunity.
[00:44:07] But we wanna help you to time it the right way. So one of the things that we're thinking a lot about is how are we going to start to get into those answers? So you see this title, optimizing your website with Tinder loos, SNCC links, and BRI tags. Obviously, what's a snicker or not a snicker doodle, a trend loo, a SNCC links and a BRI tag. [00:44:27] There are words we made up. We are testing. How long does it take? How little work or how much work do we have to do to show up in those answers? And we are going to continue to blow this out. If you are someone that's yeah, I wanna stay with you guys on this, then that is Anthony who is now running all of our product work and all of our engineering work. [00:44:48] And he's starting to work with
one of our engineers who's now out on paternity leave to start running these tests. So if you're like, yeah, seer, like I wanna know. The tests you guys are running and how you're getting into [00:45:00] chat, then let us know, because then I can have him expose more of the work that we're doing in a follow up webinar. [00:45:05] And again, that's why I really wanna be partners with you guys in this process because I need to know what things you are most looking for and how we can tackle that. So I'm gonna wrap up right now, and then we're gonna get to some questions. So it's what should
I do? What can I do? One, I'm changing the way I interview. [00:45:22] If I could do it, I don't think, I dunno if it's legal or not, so we'll have to talk to the lawyers. I'd be like, yeah, show me your Chad g p t history because it's so new that if I'm gonna hire you, like I wanna know you're in the tools. Because I don't know all this stuff and I'll tell you, but I sit with somebody and they're in it. [00:45:36] And I'm in it. The stuff we come up with together is crazy. Like I took a little thing and then Jordan made it
a thousand times better. You got to find the people in your organization who are leaning into this, and if you're doing some hiring right now, it's a great time to start to stack the deck with people who will help you to be better. [00:45:55] I'm not gonna say, take a team member and take google.com off their pc. I'm not gonna go that far, [00:46:00] but you gotta find somebody who's gonna be your Sherpa internally, somebody that's pressing and saying I don't Google things anymore because I want to guide our company in a way where people start to know I'm the guy or the girl to go to when people have these questions. [00:46:16] And you better believe CEOs, whatnot. They have these questions. Why? Cuz Wall Street is going, what are you
doing with ai? You're not mentioning it in your quarterly reports. If you don't have any buy-in, so freaking what, I'll be like, oh, how can I get buy-in and my company won't let me do it. [00:46:28] If you have a personal computer and 10 bucks a month, there's no reason for you to not be playing with this because there's a chance that if the companies don't eventually get on this. Train, they're gonna be on the tracks and they're gonna get run over. And I don't want any of my clients at least to be behind the times. [00:46:46] So get sharp at this. So even if your company never fully adopts it,
you are sharp enough to go into your next interview and talk about how you are playing with chat g p t, and how you understand how Bard is or isn't working. Bard's free, perplexity free. [00:47:00] The barrier to entry is desire. And and then if you do have corporate buy-in, then work with those people to share data with your partners, not just seer other partners. [00:47:09] Cause then you also get to share with those partners and their AI teams and their AI task forces. If I gave you my chat logs, what would you do? Oh, here's how we would use it, is what you want to hear. If it's, I don't know, I drop in chat g p t and write some basic ad
copy, you're like not leading, right? [00:47:25] So I wanna, I want you to bring those kind of questions to me and the team that I'm going to build over the next year or so, and the investments I'm gonna make in ai. Bring those questions to us so we can show you how we might be able to use that data plus our library to get more outta chat g p t to help you all to win. [00:47:39] And with that said, Marissa, I'm all wrapped up here so we can now go to any q and a or anything that was dropped in the chat. Fantastic. We got a
few in the chat. And if we have some time, I will hit on a few that came through prior to the event as well. So first question is around hallucinations. [00:47:57] So would hallucinations be mitigated through the use of [00:48:00] appropriate plugins because a basis answers off of a realtime response from an a api, for example, travel plugin for related travel queries? That's a really good question. If they had plugin to Google. Which I don't know if they do or not. So
I can look into that or share with me. [00:48:17] If you do. Cause we're gonna make each other better here, remember? I would wanna know that because I would wanna Google it. So when you go in a bard this is the beauty of having some alumni that work at Google is early on I was seeing how it worked and they were like sharing stuff with me.
[00:48:29] And it's look like, Hey, here's what it says about you will. And I'm like, wait, that's wrong. That's wrong, right? And then I would Google it and be like, does Google have the right rich snippet? Answer? So remember, we index all those answer boxes on Google. This is it.
[00:48:44] This is the partnership I could take an answer from for a question. Then I could go and pull all the people Also asks where Google typically has a better answer, sometimes with no hallucination. And then say, how close are the, I could use chat G P T to say, how close is [00:49:00] your answer to the answer that Google has as more of a fact inside of their answer box. [00:49:05] And because I have all the answer boxes and all the data for the answer box in my library, we could literally do that. If it creates value for clients, we could do that and start working on that next week if it created enough value for us to put our time there. Awesome. All right, another question.
So LLMs get dinged for biases that we know exist in text, such as racism, sexism, elitism, et cetera. [00:49:28] How do we mitigate against that? I dunno. Is, I think ethicists, ai ethicists that we need to follow and learn from. My objective is to help my clients. To beat their competitors and to win. So for instance, when I showed the black doctor example that's happening today,
people are searching for black doctors and we're dropping them on pages with white doctors. [00:49:59] That's not culturally [00:50:00] relevant, right? And that's not about ai. It's I wanna find where we're just making friction for customers. So I can probably get you a list of a few of the AI ethicists that I've been following, but it's a huge issue. It is a huge issue. One of the things that one of them found was that like self-driving cars in Vegas, I think we're more likely to hit black. [00:50:18] People because their training set
wasn't trained to see them as much as like a human walking across the street. These are real problems and I've talked about it. I've also blogged about, and we can share it later about some of the issues with even the image ai. It sees a black woman in a doctor's uniform and it thinks that matches to the word streetwear.
[00:50:36] So there are a lot of issues there and we are playing with it a bit and finding and trying to surface these things to send back to Google and other companies. But I don't really have an approach on that just yet. I'm really focused on how to do, how to create a competitive advantage for our clients, and then I can connect people to folks that I know are looking at the AI ethicist side. [00:50:54] Awesome. All right, another question here. So if chat sheet PT has only been [00:51:00] trained since 2021, how can I be sure that changes that I'm making are more visible? To be more visible, have an effect on it before they retr
2023-06-27 16:13
