Revolutionizing Small Publisher Advertising Through AI

Chris Kiefer (00:00.726)

Welcome back to The Pursuit of Purpose. My name is Chris Kiefer, and today I am joined by Kenny Katzgrau, who is the CEO of Broad Street. So I'm going to first thank you, Kenny, for coming on.

Kenny Katzgrau (00:13.267)

Yeah, thank you for having me, Chris.

Chris Kiefer (00:14.918)

and give us the brief rundown of who Kenny is and what Broad Street does.

Kenny Katzgrau (00:20.894)

Yeah, well, Kenny is a software developer turned entrepreneur or SaaS owner turned news publisher. I think since I was a kid, I wanted to start a company. And through many iterations and many failures, I eventually arrived at Broad Street in 2012, which I had been growing probably for like four or five years as a solopreneur, just grinding it out to make it happen. And then, uh.

for another six years today with a team of like 13 full time. And then, you know, a lot of contractors to support that. But Broad Street is a very like very purpose driven company. Actually, we stand for small publishers and a lot of people are not unaware of the, the decline in advertising revenue in the news industry and the publishing industry in general.

A lot of that ad revenue has gone to other places, primarily Facebook and Google. So when I give my perfect intro, as I say, like, hi, I'm Kenny Castor, I'm the CEO of Broad Street. I usually say we help small publishers kick the shit out of Google and Facebook. Um, it is. It is like our purpose for existing is making sure that advertising with a small publisher is a better option for that business, a lot of different ways to make that happen, then spending money.

Chris Kiefer (01:33.403)

Hahaha.

Kenny Katzgrau (01:47.19)

boosting a Facebook post or running Google ads is for that business. So I want to make sure that advertising and marketing with a small business is actually the best option. And we do that not just to build a company. I didn't get in this to make money, because if I did, I probably wouldn't have picked something that took me five years to actually make a decent profit. We do it for the small publishers across the country.

Chris Kiefer (01:51.599)

Mmm.

Kenny Katzgrau (02:14.626)

The ones who have written to us and said, hey, man, we wouldn't even be in business if it wasn't for you guys. And because we know that they are doing something great. They are shining a light on what is going on in their community. Sometimes it's on the good things that happen. And sometimes it's on the things that people need to know about what's going on in local government, what's going on, you know, with, with like real estate, what's going on. They are shining a light and really you cannot have a healthy democracy.

if you don't have the journalists and reporters of the world working to shine a light in dark places. So that's why we really do it. We work with like 500 news publishers and that's fantastic. But I know by extension, we help keep those publishers in business and shining a light in their communities. It extends to millions and we work primarily in the US now, but have been slowly and accidentally expanding globally.

Chris Kiefer (02:50.045)

Mmm.

Chris Kiefer (03:02.162)

Mm.

Chris Kiefer (03:08.718)

That's awesome. So I'm super excited about today's conversation because we cross-passed through SAS Academy and I saw a post that you made on ChatGPT and it was a presentation that you had given. But when we met a couple weeks ago, I feel like this, I love, love conversations around AI emerging technology. How do you, like, first of all,

Kenny Katzgrau (03:17.452)

Mm-hmm.

Chris Kiefer (03:38.51)

what are the use cases for it? Some of them just kind of silly party tricks, other things, very, very practical time-saving things in the business. And I think I wanna start out with a phrase or a quote that I heard from Dan Sullivan, which is, in 10 years, there's gonna be two types of businesses, the businesses that are using AI in their daily operations and the businesses going out of business. And I personally believe that.

Kenny Katzgrau (03:47.158)

Mm-hmm.

Chris Kiefer (04:06.734)

just from my experience, I've been on a journey to connect with people like you and other people that are actually trying to utilize it. I feel like there's a lot of, um, fear around AI and taking over the world and, um, you know, the bad things that are going to happen or how people are going to misuse it. But I feel like personally I choose, like I am choosing to figure out how we use it for good. Cause yes,

Bad people have always used everything in the world for bad, but there's always the heroes that are combating the negative or the ways that can be used negatively to positively influence and build a better life and world for everybody. So that's my frame leading into this conversation. I don't know if you have any opening remarks, but I'm super pumped.

Kenny Katzgrau (04:37.558)

Mm-hmm.

Kenny Katzgrau (04:54.014)

I think, yeah, I think you touched on something pretty deep. There's this eternal battle between good and evil. And I think the optimists of the world would say that good will win. Although I think movies like Terminator obviously set us up for the dystopian outlook for the future. But I don't think that's necessarily the case. It's kind of like, it's like with computers and mobile phones. I don't think that there's anything that is necessarily bad. I think there are trade-offs, certainly.

Chris Kiefer (04:59.895)

Yes.

Kenny Katzgrau (05:24.21)

Um, so, you know, the fact that we can't really, you know, grab dinner with our significant other without checking our phone is maybe one of the drawbacks of having a computer in your pocket all times, but at the same time, it can be certainly like handy, you know, when your kid has a phone and they're off hanging with their friends, you kind of know where they are, can get in touch with them. So there's that too. Um, I think there are trade-offs and I think finding the good in all change is probably not a bad way to do it.

Chris Kiefer (05:53.742)

Yeah, and I feel like there's another perspective of like, I wish we could just go back to the 1800s or whatever, people have said or say things like that. And I feel like, okay, have fun wishing. It's like the world is moving in the direction that it's moving and it's generally like going towards ease and more comfort and all that stuff, which you have to be mindful of. But I feel like

Kenny Katzgrau (06:02.21)

Mm-hmm.

Kenny Katzgrau (06:07.958)

Yeah.

Chris Kiefer (06:23.234)

You just have to acknowledge where we're heading and then decide, are you going to be someone that makes an impact and guides or directs the direction? But you're not going to just like turn the train around and go in the opposite direction, you know.

Kenny Katzgrau (06:56.164)

Right. Now

Kenny Katzgrau (07:06.63)

and wonder of it all? Or do you want to focus on all the negatives and wish for a time that honestly you've never experienced yourself? And if you are the kind of person who focuses on the negative, we transport you back to 1850. You're probably going to be focusing on the negatives there too. So focus on the positive.

Chris Kiefer (07:23.211)

True, love it. Okay, so I don't know how you, how do you wanna do this? I know you prepped some use cases, but what I'm, correct me if I'm wrong, we're gonna try and do in the next 30 minutes is give you, what was the number, six or seven practical applications of how you can take AI and use this in your daily life or in your business like today, is that right?

Kenny Katzgrau (07:46.526)

Right. Yeah, definitely. And I'll kind of go through, yeah, it kind of starts with more or less like the party tricks, but then it gets into stuff that's like really, really time-saving and then I'll kind of end on my favorite. Right. So the one that I think, the one that like everyone tends to gravitate toward, at least initially, is this concept of like idea generation.

Chris Kiefer (07:59.163)

Okay, sweet. Well, let's hit it.

Kenny Katzgrau (08:16.042)

Right? And for anybody, and I'm not sure how many listeners might be familiar with SAS Academy, but.

Kenny Katzgrau (08:25.898)

whether or I'm actually in publishing and a lot of times it's about like generating ideas for stories or like you know if you're an entrepreneur and you want to write blog posts like generating ideas is like a natural place to go to chat gbt personally in practice I don't find it to be like necessarily generating like the best ideas ever but um uh every now and then if like it generates two i or 10 ideas two of them are like really usable so you're an entrepreneur and you know you

Chris Kiefer (08:52.742)

Mm.

Kenny Katzgrau (08:55.394)

thinking about blog posts, what are some of the things I should write about, related to your industry, like of course you can ask ChatGVT. What I find ChatGVT more useful for is that when you have an idea for a blog post that you really want to execute on, sometimes asking it to help support some of your assertions. I guess you can have it like make an outline. If I were to say, why...

Chris Kiefer (09:18.194)

Mm.

Kenny Katzgrau (09:22.73)

This is just an example why the future is moving toward digital advertising. Right. Um, chat, GBT can certainly make, uh, like an outline for me, but a lot of times I find that like, if I can really kind of write the post, but use chat, GBT to help me find like kind of, uh, supporting arguments in favor of something that can be really powerful. Because sometimes in the process of writing a post and like, you're, you're like, wait, do I have any evidence to really support this? Are there any.

examples, counter examples, or something I could use to support my assertions here, I ask ChatGPT a lot of the time while I'm writing those blog posts. So idea generation, but really generation of ideas in support of the thing that the content that you're actually trying to produce. So on the marketing side, I've used it a lot for that case. And of course, like, you know, just to give you, once you write that blog post,

Chris Kiefer (10:14.063)

Hmm

Kenny Katzgrau (10:22.026)

writing like 140 or 280 character summary, you can paste it all in the ChatGBT and basically say, give me an X or Twitter post for this, a Facebook post, et cetera. So these are like, this is like ChatGBT 101 that pretty much everyone uses it for.

Chris Kiefer (10:41.622)

Yeah, I just, uh, while you were talking, I went in and pulled, I just type that in. I'm writing an article on why the future is moving towards digital advertising. Can you provide some evidence to support examples, counter examples? And it says, I'm not going to read the whole thing, but video content dominance with a, just with explanation, personalization at scale, blockchain and marketing, augmented reality and virtual reality, inclusive marketing.

data strategy, contextual advertising, on-site personalization through dynamic creative optimization, in-house marketing operations. And each one of those has, you know, I feel like to your point, you'd probably skim that and be like, the personalization at scale or something is like really jumps out like, ooh, that's, if you can have more targeted ads, more personal ads, that's obviously gonna be more useful, but yeah.

Kenny Katzgrau (11:36.394)

Oh, totally. And yeah, exactly. So I would kind of cherry pick the things that I kind of agree with. You know, getting into like augmented reality and virtual reality, that's like, that's too far off to really speak intelligently on. So, but the personalization, absolutely. The ability to target, of course, you know? So, so sometimes like what it comes back with can kind of spur more ideas and kind of help you develop the concept a little further. One of the things that like,

Chris Kiefer (11:41.371)

Mm-hmm.

Chris Kiefer (12:01.414)

Totally.

Kenny Katzgrau (12:04.106)

I don't know if anyone's ever had this experience. Sometimes it's easier to edit something that already exists than start from scratch. So a lot of times if I'm having my team put together a newsletter for the next month that goes out for our company, they will take the first crack at it and I will just zoom in and I'll see even if it's the complete opposite of what I wanted to say.

Like it kind of triggers me to write the thing that it like more effectively brings that out of me. Um, so a chat, GBT, like if it's something that you agree with or disagree with, just having the starting point. That's something that you can kind of, um, you know, exercise your intellect against, um, is a great tool because you're not starting, it's not just like tumbleweeds in your head. You have something to start.

Chris Kiefer (12:42.179)

Mmm.

Chris Kiefer (12:48.391)

Mmm.

Chris Kiefer (12:53.274)

It reminds me of, I would do, my early career was video production and we do these testimonial videos or client interviews for our clients, right? So the clients of our clients and you know, there is a skill and a craft in getting them to say something that's usable for the video and my favorite was filling the blanks like statements, which I feel like is very analogous to what you're saying. So it's like, Kenny, I want you to say

Kenny Katzgrau (13:17.932)

Mm-hmm.

Chris Kiefer (13:20.09)

The thing that surprised me the most about working with Boolean was, and then you say on camera, the thing that surprised me, and then you just are off and running, you know? But oftentimes I feel like I had been in interviews where they're just like, what do you want me to say? And it's the, if you just have, just like, get me started. Start the car rolling down the hill and then I got it. You know what I mean?

Kenny Katzgrau (13:29.004)

Right.

Kenny Katzgrau (13:41.45)

Right. Oh, totally. Yeah, that is the exact same thing. So on the topic of idea generation, and this is where I find it a little more helpful. Because as an entrepreneur, I'm very focused on how I can build a very finely tuned, effective organization. And I'm focused on all the little different pieces. And SaaS Academy has been a big influence in that. And there are some people that, even Dan in SaaS Academy,

have been influenced by. But something that SaaS Academy is very much about is developing simple systems to remember things. So there will be a lot of acronyms. So it'll be like the smart system for sales. And like S-M-A-R-T will be acronyms. And Dan has said, he's the founder of the Creative SaaS Academy. Dan has said basically, if

If there's no system, like it won't travel. And meaning that if you don't put like a concept or something into a memorable form, like it's really gonna be hard to like pass that onto somebody else and in a way that remember it, right? So a perfect example of this and something I use ChatGVT for is that I wanted to put a very simple like social media marketing plan together, right?

Chris Kiefer (14:58.226)

Hmm.

Kenny Katzgrau (15:07.906)

And I was thinking like, all right, well, first, what do I do? Like I come up with like the thing or the, the core of the content. Right. So I do a lot of webinars. I do a lot of blog posts. I do there a lot. There's a lot of Broad Street, Kenny Castro content out there. So like I'm originating the content, but then what I need someone to do is kind of like, go through all of that content, sift through it and find stuff that would make a good, uh, you know, baseline for a social media post.

If you go back and broadsheet social media, there are years that went by when we had nobody posting the Twitter or Facebook, anything. So I wanted to solve this problem, make it a really simple system. So I'm like, first I originate the content, then I have someone who sifts through the stuff I've already done, just finds all the smart stuff I said and interesting things we've done, so they get that. Then I have someone who should share it. This is just the person who takes the content and puts it on social media.

And then sometimes that stuff is really good. It performs well. And I should have someone who boosts the post. Right. So basically I called this, I was like, whatever this is, I called it my bomb marketing plan, right. And it was so perfect. Cause I was asking chat, she BT to like what I would give it a word that I could use and I would say like, this is my process, give me like an acronym that could describe this process effectively. Right.

Chris Kiefer (16:16.55)

Hmm

Kenny Katzgrau (16:32.958)

So what it came up with was Fuse, F-U-S-E, which is perfect for the bottom marketing plan, right? So, formulate, which is me originating, uncover, which is someone on our team uncovering the good content, right? And then putting it into a spreadsheet, share, which is someone looking at that spreadsheet and loading it into HubSpot or TweetDac or Hootsuite or whatever they use now, Buffer. And then elevate, which is our paid marketing person,

Chris Kiefer (16:39.492)

Ah.

Kenny Katzgrau (17:02.098)

swooping in to see what performed most effectively and boosting that. So FUSE made that acronym with the help of, with ChatGBT. So a lot of people I've told that to, and they're like, that's awesome. Can you send me that document? So I have it all in a Google doc, FUSE four pages, and everyone remembers it.

Chris Kiefer (17:06.414)

Mmm.

Chris Kiefer (17:22.49)

But that's a, I think the thing that's great about that is again, like Dan said, people, the reason people love it is because it's not anything they haven't heard before. Like everything you just said is like, yeah, obviously, but it's the packaging of the idea in a memorable acronym. That's like, that's cool. Bomb marketing. It's a fuse. We've got to get, I got to light the fuse. Let's go. You know, that's awesome. Would you put that, is that still under, are we still under the, uh, idea generation, uh, piece?

Kenny Katzgrau (17:31.594)

Mm-hmm.

Kenny Katzgrau (17:44.172)

Right.

Chris Kiefer (17:51.182)

of our seven points here. Okay, so in the essence of time, I'm trying to keep an eye. We've got like three minutes per point in order to get all seven. So what is number two?

Kenny Katzgrau (17:51.638)

Yeah.

Kenny Katzgrau (17:58.626)

Okay.

Kenny Katzgrau (18:02.474)

All right, so number two, this is a little more technical. But, oh well, so that was actually the second version. So the first part is just generating the ideas. The second instance is actually using that idea generation for something a little more practical, which is not just your standard block content, but refining it and also this is like, it's for internal processes, making sure that, so we're kind of moving categories here. One's marketing, one's actually operations.

Chris Kiefer (18:18.65)

Refining it. Yeah.

Chris Kiefer (18:32.11)

Hmm. Yep.

Kenny Katzgrau (18:32.258)

for your company. So just like now that we're kind of in the operations piece, because I like mentally, I think of like marketing, sales, operations, product, or finance, whatever. So I kind of move through that parts of the company. But the next part is basically as a developer, someone as a software developer, and this is for the technical folks out there, there are a lot of times that you know how to do something.

that ChatGBT can just do it faster. So like you need to write some code that would probably, yes, I could Google it, I could go to Stack Overflow, start to test some code and put something together. But let's just say I need a method or a function, or I need a method or a function that does something that I can ask ChatGBT to do. So a perfect example is that I was writing a WordPress plugin and I asked ChatGBT, write me a boilerplate WordPress plugin, which it did.

Chris Kiefer (19:05.443)

Hmm.

Kenny Katzgrau (19:31.754)

And I said, make it so that when a API request comes in from Twilio, it takes the request and then stores that in the database. So I asked ChatJPT to do that. It modified the code, and it wrote all the code to do that. There were a few bugs along the way, which were really annoying. But ChatJPT wrote that all out. Then what I did was I said, how about when it gets the request, which was a text message from an end user?

take that and make an API call to Chat GPT and have it correct for spelling and grammar. So basically, I just kept telling Chat GPT to modify this code in this WordPress plugin. It was writing over time. And eventually, after some debugging, arrived at a WordPress plugin where Chat GPT was basically making a call to its own API to do the things that I needed it to do. Now.

The reason that was fantastic is because I didn't know what ChatGBT's API signatures were. I did not want to use a giant code library as part of this project. I wanted to keep everything in a single file. And ChatGBT made that possible. So whether it's writing code, whether it's writing queries for a database, if you're a developer and you're still Googling it and going to Stack Overflow, ChatGBT should be number one. It should be the first thing that you

Chris Kiefer (20:36.017)

Hmm.

Chris Kiefer (20:49.873)

Mm.

Do you use the copilot in... Oh man, I'm spacing. My developers, in what?

Kenny Katzgrau (20:59.478)

Visual Studio Code.

Chris Kiefer (21:04.274)

I can't remember what the name of the website is though. What's the...

GitHub, GitHub Copilot.

Kenny Katzgrau (21:11.454)

Yeah. Oh, yeah. Yeah. So, um, so right. GitHub, co-pilot, and then the, uh, code editor that I think a lot of people use is visual studio code. And, um, so I use co-pilot within that and, uh, yeah, it's really helpful. Um, where it can kind of take you the wrong way. Cause a lot of people know that chat GBT ain't always right. In fact, it's not right a lot of the time is that occasionally it writes bugs for you too.

Chris Kiefer (21:21.635)

you

Mm.

Yeah, okay.

Kenny Katzgrau (21:38.73)

So if you don't understand exactly what it's doing, there may be hidden little bugs in there that you have to address. Right? So, really, that's number three. Right? All right, so number four is actually something that is on the sales end, right? And this is something that, there are a few different ways to do it, but like this is,

Chris Kiefer (21:39.131)

Hmm.

Chris Kiefer (21:45.382)

can come back and bite you, yeah.

Chris Kiefer (21:49.646)

So that's number two, or is that number three? Three, okay, and then number four.

Kenny Katzgrau (22:10.91)

It's, you can, if you have a sales conversation and you're recording it with something like Otter or Fathom, right? These are like different, or Firefly AI, they will hop on your Zoom call with you and summarize the Zoom call and give you like the primary points. And that's like a fantastic use case. But something that I did the other day, I was working with somebody that I needed to, I needed to kind of interview them. I needed to get like answers to three questions.

Chris Kiefer (22:20.827)

Mm-hmm.

Kenny Katzgrau (22:40.118)

because I wanted to write content around those three questions. So I had originally tried to do this by email, right? I had sent this person three questions that I asked them to answer, and they really just kinda didn't do a great job in giving me their response. They kinda gave me one-liners for each. And I know this is a very talkative and gregarious individual. So I said, you know what I'm gonna do? I'm gonna get them on the phone, literally like old school phone.

And what I'm going to do is I'm going to record the conversation and I'm just going to kind of get them into that mode where they're just talking. It's exactly what we were talking about before where you kind of have to spark the conversation. So I asked him these three questions. In this case, it was like a real estate agent. And I was like, all right, yeah, so we just bumped into each other on the street. I'm like, hey, what's going on with real estate in this part of, you know, Red Bank where we live? And he went on for probably a good like five, 10 minutes, had so much to say. But before he just gave me a one liner.

So like, all right, fantastic. Next question, after about 40 minutes, I have tons of content. Now, who's gonna sit through and like, listen to this call again, and now I gotta type it all in and make the content? So what I did was I took the audio file and I uploaded it to Otter AI, right? And Otter did the transcription, it said person one, person two, it was like perfect. And then what I did was I took that to ChatGBT,

Chris Kiefer (23:37.714)

Hmm.

Chris Kiefer (23:47.77)

Listen to that and try and distill it, yeah.

Kenny Katzgrau (24:08.03)

This is a conversation about so-and-so. Here are the three questions that I asked. Using the audio transcript, write brief answers to each of those questions based on the audio transcript, right, and just paste it in. And ChatGBT basically turned what would have been like four or five hours of work into like 10 minutes. It was amazing. So, and that like.

Chris Kiefer (24:22.578)

Hmm.

Chris Kiefer (24:29.25)

Yep. I love, yeah. I was gonna say that is to me, this is probably one of the most useful use cases is where you have the outline and then you've got a lot of just like ancillary or like just content from like this conversation, this conversation, this conversation. And then you say, like, I wanna build on what you just said. This was a random example, but we have this house that my wife and I are interested in buying.

Kenny Katzgrau (24:58.242)

Mm-hmm.

Chris Kiefer (24:58.33)

And I called the owner of the house and I had a conversation with him and understood like what he was trying to do. Cause I was trying to do seller financing, just different unique situation. So I had this conversation with him. That was like 10 or 15 minutes long. Then I also got the Zillow information. And then I said, here are my goals and objectives for the house. And basically said, can you draft up a letter that I could send to this guy?

Kenny Katzgrau (25:18.496)

Mm.

Chris Kiefer (25:23.374)

And I wanted to include things like purpose and what is the, like what was his legacy that he wanted all this like, like ethereal or like, um, just, it was a complex, like a lot of different stuff. And it writes this beautiful letter of like, Hey, I know you mentioned this and you know, and it pulls in this information from all these different things. And then you can obviously be like, make it a little bit like less professional, like use simpler language. And then it rewrites, you know, something that would sound like me or whatever. But anyways.

I think that like sync order, is it syndication or synchronization or what's the word I'm looking for? Synthesizing of information of just like a lot of stuff from multiple places and boil it down into like this output.

Kenny Katzgrau (26:01.102)

Mm-hmm. Right.

Kenny Katzgrau (26:09.15)

Oh, totally. Yeah, so just two notes on that. So something that I've done in the past is very simple agreements, especially for someone that something's not going to get litigious over, like just a short agreement with a freelancer. So short contracts. I actually was a recipient of a subpoena from Google and the Department of Justice over the summer in the federal antitrust thing. So we got swept up in that. And I asked ChatGVT, I was like, write a response to the subpoena. And it did.

And I was telling one of my advisors, and he was like, do not use what ChatGPT. He wanted me to go with the lawyer, which I did. But it actually did a fantastic job writing the response. It was a simplified version of what the lawyer actually did. But there are a lot of great use cases like that. Because if you ask it, I have another friend who's a social worker. And for their cases, they have to be formatted a certain way when they take it to court.

Chris Kiefer (26:39.046)

That's awesome.

Chris Kiefer (26:53.394)

The lawyer did, yeah.

Kenny Katzgrau (27:07.69)

So what they do now is, or what he does, is that he takes all the facts of the case and asks Chachi Buti to basically write a report. The standard is really well, is already well known. And then Chachi Buti knows it, and then just takes all this facts and writes it in the report. And has received some commendation from the judge in these cases because of how well everything was formatted. It's crazy.

Chris Kiefer (27:32.882)

All right, number five.

Kenny Katzgrau (27:34.95)

Okay, number five is one of my favorites, and this is like one of the highest leverage items for any entrepreneurs out there, or really if you're in a company and you're in management or something like that. But basically, for process documentation, everyone knows it's like kinda like something you gotta do, especially for the key processes in a company. What I have been doing is that I will just verbally dictate how a process works, so like our sales process. It's got like a bunch of different steps.

And I added a whole lot of details to how each step was performed. But I verbally dictate into the recorder on my phone. And then what I do is I send it over to Chat GPT and ask it, I tell it that this is a process documentation, verbally dictated, and to break it down into steps that I can copy and paste into a Google Doc. Now keep in mind, this was a five-minute long video or recording.

I made all kinds of mistakes. I said like, uh, you know, um, I was like, all right, scratch that, erase that last sentence, you know, because it's hard to get it perfect. So, and when I got the transcript from the phone, there was zero punctuation. I'm like, well, let's do it. And chat GBT nailed it. It nailed it. It gave me a 12 step process that by the way, I just hired a new salesperson this year, just came on in January and that's the documentation.

Chris Kiefer (28:46.116)

Yeah.

Kenny Katzgrau (28:59.318)

that they went over. Of course, I corrected a few things along the way, but the reason that's so powerful is that a lot of times as business owners, we don't get an opportunity to document the process. So what we end up doing is spending our time repeating ourselves to different people that join the team. So with this written document, I have leverage now that any new salesperson who joins a team, I can just give them that document. Importantly though,

Now my team is doing the same thing. Now they're the ones verbally dictating and creating process documentation. So we took the process of creating documentation and like really drove down the total time investment required. And in the process, we're actually spending far less time repeating ourselves and telling each other what to do. Right? So basically, any time you're getting more work done with less effort, that's called leverage. And

Chris Kiefer (29:48.274)

Hmm.

Chris Kiefer (29:55.472)

Right.

Kenny Katzgrau (29:55.778)

you know, building a successful business is all about leverage. That's the key. So that's one of my favorite use cases because I use that probably like a few times a week.

Chris Kiefer (30:00.355)

Mm.

Chris Kiefer (30:06.966)

I love that. And that's a super practical one for like, how often do you, um, like for me, I just sitting down in front of the computer to write out step by step or to document what I do is much, is just painful and hard sometimes, but just, instead just as I'm driving around or on the way back from the gym, be like, all right, what do I do? And you know, it's just the ability to use your voice when you're not in front of a computer to.

start the process and then just refine it when you get back to the computer is great. Number six.

Kenny Katzgrau (30:38.442)

Absolutely. So last, yeah, so this is the last one. And this is probably, this is a little more technical, but it's something that you have to know about. So we've been talking a lot about chat GPT, right? But occasionally you run into a limit of like how much you can send to chat GPT, right? So a good example is like something that we do is NPS scores, net promoter scores. So basically of all your users, like,

What's their rating out of 10? You ask them some questions. And so you have tons of, especially if you have hundreds of responses, you have hundreds of responses. Where do you even begin? You have the quantitative data of how people rated you. But when you have the qualitative feedback, how do you go over 400 responses and get a general sense of what the top five priorities for the next quarter should be? You don't, right?

Chris Kiefer (31:23.09)

Hmm

Chris Kiefer (31:31.184)

Yeah.

Kenny Katzgrau (31:33.066)

Now, if you can feed them all in the Chat GBT, this is another great use case. You can have Chat GBT find like what are the five things that people seem to be talking about the most, right? But whether it's something like that, or maybe if you have like a knowledge base that you would love, imagine Chat GBT could consume everything and you give it your entire knowledge base, right? And then your customer support reps or your customers,

Chris Kiefer (31:41.138)

Common threads, yeah.

Kenny Katzgrau (32:02.67)

could basically ask a question that ChatGBT could deduce from your knowledge base. So this is called retrieval augmented generation. And it's basically like you've heard build your own chatbot. There's a framework called Llama index that a lot of developers use. It's a Python framework. There's another one called Lang chain. And you can consume all of your own data. And what it does is it creates these things called embeddings

um, for different framework, uh, for different providers. But what you can do is basically work with, uh, open AI, the underlying engine for chat GPT to, uh, basically ask questions about your own data. Right. So if you have, you know, megabytes and megabytes of text feedback from customers, what you can do is use Lama index to consume all of that and then query against it, basically build your own chat bot. So the way we're using this internally.

Chris Kiefer (32:50.076)

Mmm.

Kenny Katzgrau (33:03.79)

is that we have consumed all of our knowledge base. We've consumed all of our webinars, all of our PDF documentation, all of our help scout, our Zendesk-y responses to customers. And at this point, we have in a robot what is essentially like our smartest and most capable employee, right? Because now you can ask it a question and it is remarkably accurate in terms of what we would say to our customers. So...

Chris Kiefer (33:24.614)

Mmm.

Chris Kiefer (33:31.622)

Hmm. Do you think that it's like in that situation? Could I? Something that I've thought about is that I have all these emails to and from customers over the last 18 months of little questions Bugs things so like hey, this isn't working Is it to the point where you could just like I don't even know if you can export all emails from Gmail but just be like here's a bunch of emails from the last year and Take out the problem the like key

Kenny Katzgrau (33:33.148)

Uh.

Chris Kiefer (34:00.054)

issues that people, customers had, and then document our responses so that next time a question comes in, it can go reference the database of what we've used in the past to resolve that issue. Is that possible?

Kenny Katzgrau (34:13.126)

So the way I see it playing out is not even having to export it. And I was actually speaking to a friend of mine. He was about to finish up his PhD and has a startup called a mysterious. And he was giving me a demo and basically it mysterious. It starts out as a Chrome plugin, but you can connect it to your Gmail. And it does exactly that. So not only can you ask it questions like that, like what are the most common threads when you get an email from somebody.

you can have it suggest a response, right? Yeah, so that's probably the way it's gonna play out. And the way my friend Shamoon at Mysterion is gonna do it is that he also wants to plug in not just Gmail, but also into Zendesk and different places because as a company, your stuff is everywhere. You've got stuff in HubSpot, you got stuff in Google Drive, you got stuff all over the place, but all of that is important. Have you ever had that?

Chris Kiefer (34:44.858)

Nice, yep.

Kenny Katzgrau (35:09.598)

moment where you're like, where did we have that spreadsheet? I know we had that, but like you're looking through your email, you can't find it in Google drive. Like, you know, there are real problems that AI could solve. And like, that's another thing you spend 30 minutes looking for that spreadsheet. Um, or maybe an hour or two hours. And like, there's some tool in the future that could kind of resurface that within a couple of seconds. That's really powerful. So, um,

Chris Kiefer (35:13.848)

Yeah.

Chris Kiefer (35:26.386)

Hmm.

Chris Kiefer (35:32.846)

Yeah. And I just think that to me, people like, there's probably someone that's like, Oh, it's not, how is that going to be accurate? It's like, just the fact that you could have, like, I have a, um, an executive assistant who manages my email. And so she's trying to gain, like learn and understand how to respond to more and more and more of my emails because most of the things are like, occasionally there's something that's like, has to be Chris's response. Like,

Kenny Katzgrau (35:45.506)

Mm-hmm.

Kenny Katzgrau (35:53.825)

Mm-hmm.

Chris Kiefer (35:59.834)

but it's way less than I thought it was when I started this. Cause I was afraid to get an executive assistant cause you know, Mo my, you know, people are asking me for stuff that, and again, Dan Martell talked about like people just want an answer and they prefer that it's faster. So if it's going to take me 24 hours to get them an answer or my assistant can get them an answer in an hour, that's fantastic. You know? Um, so, but yeah, I feel like to me, it's like, if you had a draft like,

Kenny Katzgrau (36:20.083)

Mm-hmm. Yeah.

Chris Kiefer (36:27.898)

So D gets an email and it's like, here, this is what we are suggesting. She could read that and be like, it doesn't really make sense or I'm not sure about that. And then go do some additional research, which she would have had to do anyways. But how many times out of 10 is it gonna nail that, on the first guy? I mean, she's just like, and she responds to an email in 30 seconds that would have taken five minutes or 10 minutes or a half hour or whatever.

Kenny Katzgrau (36:50.018)

Exactly. Yeah. All of those really add up over time. And, um, yeah, it's like, they say like author, not editor. So, you know, I'm all about like the human review. You need that because it'll occasionally say something that's flat out incorrect. It's going to happen. I'm sure it'll get better, but I feel like there's always that risk. And at some point it's pretty much the same thing as you would expect from a human, you know, that's going to screw up occasionally.

Chris Kiefer (37:16.65)

Yeah, and that's the other thing is that I feel like people always it's like self-driving cars that Elon Musk has talked about how Like self-driving cars are literally going to have to be so perfect that uh in order to make it so that like it's most dangerous for a self-driving car when you have humans also driving if Every car was self-driving then they'd all be following standard rules because they wouldn't be you know swerving because they're drunk or

Kenny Katzgrau (37:28.418)

Mm-hmm.

Kenny Katzgrau (37:41.102)

Mm-hmm.

Chris Kiefer (37:44.93)

running red lights, like all this stuff would just not ever happen. Cause you're following like the computers are in charge, but the self driving car has to get good enough to drive with humans that make tons of errors. And then once humans are no longer allowed to drive, it'll be so like, it's like playing an easy mode all of a sudden for the cars. You know what I mean? Cause everything's now known or far less, there's far less unknowns I should say, but I feel like that's people.

Kenny Katzgrau (37:56.556)

Yes.

Chris Kiefer (38:11.47)

will frequently point out where chat GPT or AI messed up. And it's like, how many times did you mess up today? You know, like we make mistakes all the time, but we just look past human mistakes and we're just trying to find the flaws in AI.

Kenny Katzgrau (38:15.49)

Mm-hmm.

Kenny Katzgrau (38:18.733)

Oh yeah.

Kenny Katzgrau (38:27.09)

I think we're afraid that it's gonna be harder to hold someone accountable. Like, if someone crashes their car, we can just be like, all right, you're in jail, that problem's solved. But if a self-driving car crashes, it's just like, everyone's like, I don't know, there's a bug somewhere, I guess, I don't know. No one loses their job, no one so much as loses a job. So, well, probably.

Chris Kiefer (38:37.501)

Yeah.

Chris Kiefer (38:45.795)

Elon goes to jail.

Yeah, yeah, yeah. Awesome. Well, any other things from AI? I've got maybe three or four minutes for another bonus. If you got something else that's on your mind.

Kenny Katzgrau (39:01.702)

Those are the big ones. The, you know, that build your own chat bot, there are a few different services out there and there's more and more of them that will consume your own data. Just something to be aware of is that with those build your own chat bot services is that permissioning a lot of the times, like the access control list is really an unsolved issue in a lot of places. So if, for example, and I heard that IBM, for example, is working on this, like, you know, your best employee.

uh, consume all of your data with AI and then index it and, uh, make it available as like a searchable knowledge base. Well, you want to make sure that different tiers of employees have access to only the stuff they should be seeing, right? There shouldn't like, because if you have it consuming all this stuff, inevitably some of it's going to be sensitive. So how do you control and like, that's actually not an easy problem to solve.

Chris Kiefer (39:44.786)

Hmm.

Kenny Katzgrau (39:57.058)

There are real technical challenges. So it's being aware of the stuff that you are consuming. Think of who is potentially gonna be querying over that. Because right now there's not.

Chris Kiefer (39:57.298)

Hmm

Chris Kiefer (40:07.906)

Yeah, and then you have, you're going to have to take into consideration the access level of the user asking the question and have kind of a knowledge of this person has, you know, this level or like up to here, but they can't ask about wages and salaries or, you know, like the contracts language or whatever. Um, yeah, that's super interesting. Cause it would, it's gonna, if you, the ideal would be, you just say,

Kenny Katzgrau (40:26.136)

You know.

Chris Kiefer (40:35.218)

here's access to everything digital that we have, our Google Drive, our emails, everything. It's just sucking up all that data. And then how do you, like, cause if you had to, if there was a huge process of indexing and tagging and like, that's not gonna ever get adopted because it's way too time consuming to get all the data incorrectly. It's gotta be able to understand what it's looking at and then understand what a given access level

Kenny Katzgrau (40:44.247)

Mm-hmm.

Chris Kiefer (41:05.55)

can receive, you know.

Kenny Katzgrau (41:07.426)

Right? Something that is not as kind of peripherally related is that in terms of the stuff you can ask it to do or you can ask it for, you can also extend ChatGBT and or OpenAI and this retrieval augmented generation with functions. So you can basically imagine this is the use case. Imagine you're using a SaaS app like Broadstreet

Archive all advertisers that haven't been active in the last five years. If you can kind of, instead of like going through and clicking through it, as the developer, what you can do is you can specify functions that it can execute or tell you to execute if a user has asked for that specific thing. So it's kind of like taking AI and ChatGPT from like purely an interface that we think about like querying and asking it for things.

and having it actually do things, or having it actually appear to do things. And that's where we kind of get into like the Terminator thing to bring it full circle. When you start empowering it to kind of instruct to execute certain functions, then it gets interesting. But I think most SaaS apps will have some form of that. And I can't wait to host a webinar where...

Chris Kiefer (42:06.598)

and don't go do stuff.

Chris Kiefer (42:16.794)

Yeah.

Chris Kiefer (42:24.262)

Hmm

Kenny Katzgrau (42:33.07)

For us, Broadsheet's all about advertising and small publishers. When we say, our chatbot that we call Alva, which is basically Thomas Alva Edison, he's a New Jersey guy, so we named it after him, and we say, hey, Alva, create a spec ad for Surf Taco, which is like a taco place in town. And when it does that, basically executing some functions we already empowered it to do, that's going to blow minds. And

Chris Kiefer (42:53.339)

Hmm.

Kenny Katzgrau (43:01.27)

When you know how something works, the magic is removed a little bit. But for everyone else who has no idea how it works, it's going to be some real like David Blaine, David Copperfield stuff. So it'll be great.

Chris Kiefer (43:13.205)

I love that. Moving into the wrap up questions, what are your three book recommendations?

Kenny Katzgrau (43:18.598)

Oh yeah, so one of my favorites, I was just talking to Shamoon about it yesterday. And just by chance, I didn't talk to Shamoon in forever. The Black Swan by Naseem Tlaib. Some of you might be familiar with the phrase the black swan. He's a statistician, philosopher, former New York City stockbroker, really intelligent guy wrote that and a book called Any Fragile that I'm a really big fan of. A book that I'm reading right now that I absolutely love recommended to a friend, The Unwavering Power of Focus. I think the author's name is Dead Pani.

Chris Kiefer (43:27.761)

Mmm.

Kenny Katzgrau (43:48.67)

Um, somewhat like meditative, but extremely practical too. But it's about like the focus of your mind and like, you know, um, all of the applications of really being able to understand your mind and use it as the most powerful tool at your disposal. We spent a lot of time talking about artificial intelligence, but your mind is the most powerful thing that you can learn to use effectively. And, um, I'm going to have to say just like being a Jersey guy, Bruce

Kenny Katzgrau (44:18.474)

It's not all about music. And if you're a Spring Scene guy, I mean, he's up there with Bob Dylan and Joni Mitchell. Born to Run. This was recommended to me too. And I read this in Steve Jobs' biography, which is also incredible. But I probably didn't have to mention that, because almost everyone's read it. But that is a fantastic book. And if you're like a creator, if you're someone really just trying to be authentic and be yourself,

Chris Kiefer (44:24.666)

What's that book called? What's his biography called? Born to Run.

Kenny Katzgrau (44:48.474)

and you want to read a success story of how someone else did it and their journey, because it wasn't overnight for him, it's phenomenal. So.

Chris Kiefer (44:57.282)

And what was the title of that one? I wasn't looking at the screen. No, I thought you said there was a second one. Or is that it?

Kenny Katzgrau (45:00.43)

Born to run.

Kenny Katzgrau (45:05.258)

Oh, Steve Jobs biography. I meant to. Yeah, I think so. I think that's up here. Maybe I have it at the house.

Chris Kiefer (45:07.014)

Oh yeah, that's just called jobs, right?

Chris Kiefer (45:13.762)

Yeah, okay. Awesome, and what is your favorite movie?

Kenny Katzgrau (45:17.412)

Uh, Shawshank Redemption.

Chris Kiefer (45:19.534)

Oh, classic.

Kenny Katzgrau (45:21.07)

Yeah. In fact, one of our services is named after Red. Maybe it's because I'm Irish.

Chris Kiefer (45:26.258)

I love it. Yeah, you probably knew this, but Shawshank Redemption is the number one movie in the world on IMDb, highest rated. It's such a just such a great movie.

Kenny Katzgrau (45:42.609)

Oh, yeah, it was fighting the Godfather for a long time for that number one spot and Somehow Dark Knight got up. I don't know if it Dark Knight belongs on the top ten but yeah, I look at that occasionally see how the crowd

Chris Kiefer (45:46.575)

Mm-hmm.

Chris Kiefer (45:54.133)

That's funny. And what is a, if someone wants to reach out to chat with you, what's your preferred method that they do that?

Kenny Katzgrau (46:02.63)

Email, definitely. So Kenny, K-E-N-N-Y at BroadStreetAds.com. So all one word BroadStreetAds.com. That's kind of my thing. Like my EA is out on my inbox. I'm looking at my inbox all day. Even my personal email doesn't even get as much attention. So Kenny at BroadStreetAds.

Chris Kiefer (46:24.438)

Awesome. And I'm going to record this is just for happy our editor. Just cut out, cut out this part, but I'm going to wrap the episode and then I'm going to record a separate introduction based on what we talked about. So there you go. Happy. Well, Kenny, thanks so much. This has been super fun. Thanks for coming on and chatting. I love the AI discussion, and I want to get more conversations like this with other people. So.

I mean, I know you have some connections and for the listeners out there, we will have several more episodes coming up with other AI topic to AI related things, just because I believe that we have to, we have to understand it. If we're going to use it. And especially if we're going to use it for good, we, we have to understand what's possible. So thank you so much, Kenny.

Kenny Katzgrau (47:11.998)

Hey, thank you for having me, Chris. It was a fun conversation. And in terms of more people, that topic, I have some great recommendations. I'll be happy to send them along.

Chris Kiefer (47:20.452)

Awesome.

Welcome back to the Pursuit of Purpose. My name is Chris. I am here with Kenny Catscrow and we are going to be going deep into AI and giving you actual practical applications on how to use AI in your business, your personal life, creating processes, saving you time in sales calls, conversations, creating content. But Kenny has come up with six practical use cases of using AI and that's what we're gonna be getting into this episode.

So without further ado, let's get into this conversation with Kenny.

Kenny Katzgrau (47:55.31)

Thank you for having me, Chris.

Chris Kiefer (47:58.09)

Awesome, all right, here we go.