Live workshop: Riverside for coaches and course creators

00:00 Hello, to everyone tuning in. This is our very first Riverside Community Workshop, very first one.

00:08 I cannot wait to get this thing going. ah Thank you to everyone who's joining us live.

00:15 This idea really came from the idea that a lot of you in the community had reached out saying that you wanted to see and learn and hear about what other creators are doing on Riverside.

00:27 to really learn from each other. So this is kind of a jump off from our webinars. And instead, for every one of these sessions, we're going to be bringing on a different person from the community.

00:37 So you can really learn from your peers and really be able to niche down into some of these topics.

00:42 And speaking of which, we had really noticed lately that a lot of you that are using Riverside, or maybe even you aren't, and you're just joining us for the first time, in which case, hello.

00:55 um A lot of you are coaches or small business owners or course creators and you're talking all the time to clients, but maybe you're doing Zoom calls and not really getting anything out of it afterwards.

01:07 Or maybe you are doing your client calls on Riverside, but you are then it kind of stops and ends there.

01:14 There's so much that you could be doing with these client calls to help really build your business and your marketing and find new clients.

01:22 And so I have brought in the perfect person to talk to about that today. We have here Craig Shimaka.

01:29 um Craig, feel free to introduce yourself. And as he's doing so, everyone in the chat, feel free, if you're watching live, to introduce yourselves as well and let us know a little bit about you.

01:40 But Craig, take it away. Awesome. Thanks, Kendall. Appreciate it. I apologize in advance. Uh, I've lost my voice last weekend is just coming back.

01:48 So if I have to clear my throat from time to time, but, uh, I'm Craig Shoemaker and I'm a business coach for, uh, people who run online's court course businesses.

01:56 And my job is to help them build and scale out their businesses. And, uh, as I, as I mentioned it to Kendall, when we set this up, Riverside's been the foundation of my court business, uh, business from the beginning.

02:08 And so I'm just excited here to share what I have. available.

02:12 Yeah, I'm really excited to dive into it, because I did, I will say, I did get a preview, guys, of what we're going to be talking about today, and just like some really cool useful tips for all of you.

02:23 And so I wanted to start with something a little bit different, because this is a workshop, you know?

02:28 um And I wanted to, you can also answer in the chat, but I wanted to take a quick poll of the people that are here.

02:35 So um let me just get your phones out, and I'm going to ask you to... scan this QR code for me.

02:45 And what we're going to be doing is just basically learning a little bit more about the people that are in this chat.

02:52 So first, I want to ask you, I'll give you another second to scan it, or actually it'll be on the next slide.

02:59 But I wanted to ask you, hold on, play, go back to poll number one.

03:07 In just a few words, if you scan that QR code, I want to hear either in the chat or in that QR code, um what is your job title?

03:15 Just a few words. I'm also going to be putting the link to this in the, let me find the link to this.

03:26 I don't know if this link works in the chat. Okay, so we've got some financial planners, founders, podcast creators, business owners, CEOs, nice.

03:39 All right, coaches. Very cool. Wow, we've got a lot of people across the board.

03:48 And we got agency owners in the chat, business owners. Okay, cool. You guys have come to the right place. Consultants, coaches, awesome.

03:56 this is gonna, it helps us get to know what you're doing so that we can really tailor this session for all of you.

04:03 So very cool. All right, let's go to the next one, shall we? How comfortable are you with AI, like the chat bots?

04:13 Let's say ChatGPT. Are you using all the time? Are you all right, still a lot to learn, still trying to get a hang of it, or you haven't really used it at all?

04:23 Let's see what your familiarity is with all of this. And again, you can also answer in the chat.

04:28 And if you're watching the recording of this, because by the way, there will be a recording, um then just say it out loud to yourself in your chair.

04:39 That always works. us here. Awesome. And yes, so we're going to be recording this.

04:44 And we also have some really useful docs that we're going to be sending out that can share with you some of the prompts that we'll be doing. So a lot of people here still have a lot to learn.

04:54 We've got some pros. Very cool. All right. Let's keep it going. still...

05:01 And so I want to know what words come to mind when you think about AI. Are you thinking like overwhelming? Are you thinking kind of like I have to use it?

05:10 Like maybe you think about show notes and transcript first. So any words that come to mind when you think about AI, I want to get an idea of like where your head's at with this.

05:20 Cautious. That's actually really interesting. I know a lot of creators in the space are really cautious about AI.

05:26 It's actually I spoke on a panel about this during the podcast show in London that I just came back from. Excited. That's really good to see.

05:34 Unknown Wild West.

05:37 So this is very cool to see because um we're gonna be really kind of, a lot of people see Unknown Territory, Wild West, they're a little bit nervous about it.

05:50 So I think this session's gonna be great for you because everything that we're talking about today, we're gonna make it really approachable and we're also gonna show you just how powerful it can be.

05:59 Because I think that there's a misconception here that AI can like,

06:05 replaces creativity in some sort of way, but what we're gonna show you is that AI can really be a tool for you to use to really power your business and make you even more creative towards the right people.

06:17 Cool, that was fun. All right, very cool.

06:21 So that was the poll and from there, I wanna pass it over to you, Craig.

06:30 Tell me. about the process that you do that incorporates Riverside and AI, and then we're going to walk through it so that everyone here can do it themselves at home.

06:40 And again, in the chat, because it's a workshop, like feel free to uh jump in, say what you're doing, like let's have this be a conversation.

06:49 So I have some slides prepped that kind of speak to that point. Is OK if I go ahead and start sharing from there?

06:58 Excellent. OK. So I'll kind of bounce back and forth between my slide screen and ah my demo screen as well.

07:08 so, okay. Sorry, I'm just getting used to where all the windows are here.

07:20 So the big part of working as coaches is being able to have data available to you.

07:27 And it's really at the heart of just about any business. This helps reveal us as coaches or as course creators or as podcasters, our processes and the clients that we work with.

07:39 And all of this information is like worth its weight in gold, but it could disappear if you don't get a chance to really stop and capture it and use it in the right way.

07:48 So. What I'd like to do is show some AI workflows that give us a chance to use it and harness it as we go along.

07:55 uh So here I'm going to show you how to use AI models to reveal problems from your clients that you meet with, automatically find their pain points and the problems associated with them, and build a lessons library for everything that we teach.

08:13 Because I don't know how many times I've been on a call and it's like I said it just that right way or.

08:19 um dealt with a lesson in something and I was just like, if I just had that captured, how great would that be?

08:27 Riverside gives me a chance to be able to do that. Also be able to write in our own voice.

08:33 I saw in the Word Cloud, Cringy come up and absolutely that is one of the issues that we deal with when we're working with AI.

08:40 But part of the process that I want to show you is a way that you're writing. with your data and with your voice.

08:47 So even if you use an AI model to kind of help you build content, then it doesn't come across as cringy. It doesn't come across as something that you wouldn't want to read and write all on your own.

08:56 um

09:00 So being able to reveal those problems, that can help you figure out your customer's pain, unlock new products and services, and get to know their buying triggers and help deeply understand what motivates and makes them want to move.

09:15 When we have a system to catalog our lessons, content creation gets a lot easier. You're able to capture those magical insights, and you can build up your own custom IP as you're working through your content.

09:26 And all of this builds into a system that helps figure out how a system can help write like you.

09:32 So this is probably the most important side of the whole presentation. And everything that I'm going to show you is available at this location.

09:40 And Kendall also is going to be sending it out in email as well. But all the prompts and everything that I show you, you have access to there.

09:48 And I'll also show it to you again at the end of the presentation. I also just send it in the chat for anyone who's watching live. thank you.

09:57 So if I were to take a step back, I started podcasting in 2005. And not only did I have to explain to people what podcasting was, but I had to write my own RSS feeds.

10:08 And it was such a big deal, I got written up in the local paper.

10:11 But when I found tools like Riverside and Riverside specifically, I was absolutely delighted because of we have AI features available to us, we have text-based editing.

10:24 And so running my podcast or AI was such a slam dunk.

10:28 But then when I started my coaching business, I realized I have this tool here that makes it so easy for me to be able to meet with other people.

10:37 And it was just like a no-brainer for me to say, OK, well, I'm going to meet with my clients through Riverside. I can send them the recording link.

10:45 And it has everything that I need in order to give those resources to people as I had those coaching calls. But then, yeah.

10:53 I just want to say, like, I want to ask you the chat really quick.

10:57 I'm actually surprised because there are so many people that are, when I'll do community calls with people from the community, and we'll hop on a Riverside call and people will be like, my god, I didn't think to do these calls on Riverside.

11:10 So are you in the chat, are you using Riverside for your client calls? And if you haven't, have you just not thought about it?

11:17 Is there something that you are like, like, Tell me about how you're doing your client calls and if you're doing it in Riverside or where else you're doing it.

11:25 um Feel free to keep going, Craig. Yeah. And so it was, was just a natural fit.

11:31 And then I think the thing that really struck me was the fact that how good the transcripts were, because I've used different tools to try and take a recording and build a transcript.

11:41 But obviously Riverside has put so much work and innovation into making sure those transcripts are accurate so that you can do the text-based editing.

11:52 Then when I download it after the fact, it just, works really well in order to be able to say, okay, well, How can I use this to my advantage after?

12:00 Yeah, mean, I'll take like, I'll take a call in Riverside with other people that are working at Riverside. We'll record it. I'll take the transcript.

12:08 And for anyone who missed the call, I'm like, hey, I'll feed it into either use the show notes. or I'll feed it into ChatGPT and have it create some kind of notes for the meeting and send it over to the team.

12:20 Super easy. And also if there's like a really fun demo that maybe like the product team is doing, I can clip that and use it for social media.

12:27 Like there are so many things you could do after the call that I always am recording these calls and using the transcripts in cool different ways. So I'm excited to see how you're using yours.

12:36 Well, and that goes back to my original point of just the fact that there's so much data that you collect when you when you meet with people one on one.

12:43 And if you don't harness it and use it like it's, it's a missed opportunity and knowing that we have it available is such a huge deal.

12:52 So I have this slide up here for disclaimer. So I'm going to show you how to take a coaching call transcript and upload it to an AI model and do some processing.

13:01 But if it doesn't like fit your business if you have sensitive medical stuff or you know you are personal coaching get into personal stuff or financial things like learn about the process of what

13:15 I'm showing you and once we have local models available then maybe this will work well for your business but just know that this involves uploading call transcripts to the web and so you know we just want to be responsible with our clients data as we're working with it.

13:31 So The big problem is the fact that, and you can ask my wife to support this, we don't have perfect memories.

13:42 I certainly don't.

13:44 And so when we spend hours and hours on coaching calls, it's hard sometimes to keep track of what we're working with for a particular client or maybe where they're going or where the next problem that they're trying to solve.

13:56 And so by being able to extract out the transcripts and do some of this post-processing with it, It really helps you remain centered and focused as you're working through and with your clients.

14:08 And John has pointed out that we have a little bit of a locked screen because I think your notes are a little bit on the presentation. So just if you move them over there, there we go.

14:18 No, a workshop is raw, you know? This is fun. Yeah. All right, so I'm going to switch screens for a second.

14:25 And I want to show you kind of how the sort of end results of where we're going here.

14:36 And this was the part that you showed me on our first call. And I was like, wow, I really want to show this to the rest of the community. So very cool.

14:44 Okay, so I maintain basically a Google sheet of lessons and problems. Okay.

14:49 And so this is an anonymized version of the type of thing that I might collect.

14:56 And so here are, if I scroll down, you know, roughly 750 different lessons that I've cataloged from all the different coaching calls that I've had.

15:09 Now, obviously there's overlap. So I'm teaching the same things to different people at different times.

15:14 But what this gives me the opportunity to do is see how I'm explaining things differently, look for patterns, find different ways that, you know, when I'm teaching shows up for within different contexts.

15:26 And so I have a prompt that I run against the transcript after each one of my coaching calls that categorizes the lesson that I have gives it essentially like a title and then gives me a description for that lesson.

15:39 What is a lesson? thank you. Yeah. And this helps out a lot in terms of content creation.

15:45 If you want to, you know, build a, a new podcast episode. If you're going on as a guest for a podcast, this can help you quickly find essentially your IP and what you teach.

15:57 And, and the thing about the prompts is that I'm specifically telling it not to do searches.

16:03 I want to make sure I'm only using my words and only using the context of the conversation that I've had with my clients.

16:10 The other sheet that I have here is the problems that they face. And this one kind of pushes it a little bit.

16:16 So I have the problem in and of itself, like a title that I'm having it generate here, a description of the problem.

16:24 And this prompt has evolved over the months that I've been using it.

16:30 um And then I also try uh to it assign the primary emotion that my client is feeling as they work through this problem.

16:40 And so if I want to go to the sheet and I want to look at, what's all the things that worry my clients? I can easily find something like that. And then I produce a need statement.

16:49 So from their perspective, if they were to say, I need to do X, Y, and Z in order to solve this problem, what might something like that sound like?

16:58 And so it generates that for me as well. And then also a solution to where I might have an opportunity to meet the need.

17:06 And this gives me a starting point.

17:08 I may go a completely different direction than what the model suggests, but again, it's just collecting the data and giving me a chance to process it and catalog it so I can come back and use it at a later date.

17:19 And so I want to back up for a quick second and say, what would you use, maybe this a lesson, let's give an example of a lesson and a need.

17:31 Okay. So, uh, I would use a lesson.

17:34 Um, so if I wanted to create a new email series, I might go into my lessons and say, okay, well, what are some of the things that I'm talking to people about over and over again?

17:45 That seems like it's, it's just something that comes up in, in, in people I speak to want to hear about.

17:50 So I could go through the spreadsheet and I could, I could figure out and find those patterns and say, okay, I'm going to create an email sequence or I'm going to do a tutorial or something like that based off of those lessons that keep popping up on over and over again.

18:01 Or the other thing that's also really helpful is like I can go into the lessons and I can realize, okay, with that one client, there was like that moment where it really clicked for this person.

18:11 And I said something that just resonated with what they were doing. So now I can go back and revisit that and think, okay, how do I need to flesh this out?

18:20 How can I make this even better and make it something even more prominent in what I teach and how I interact with people?

18:26 Yeah, I mean, even I'm thinking about it like I do the community calls and I'm thinking I hear all the time I'll ask people what they love about Riverside and a lot of the time they'll say I love how easy it is or how simple it is for me to use.

18:37 And so I could even use that knowledge and apply that to my next email for a marketing campaign and say this next easy to use type of feature and really be able to speak directly to lessons that I've learned from these calls.

18:52 That's super cool. OK, how would you use a need? So a need it just it's the same coin just the opposite side.

19:01 So if I tend to find people are running into the same problem, or I'm essentially looking for patterns at this point.

19:08 uh And for the people that I work with, pricing and and marketing issues are always issues that come up.

19:15 And so when I'm looking at their problems, I'm looking at their the things that they're reacting to, it's like, okay, well, how can I then

19:23 figure out a way to either specifically for that client address that problem well, or then take a step back and say, okay, well, how can we sort of address this problem in an effective way?

19:36 But just like trying to remember that off the top of my head over months and months of worth of calls, I'm just gonna forget.

19:44 Yeah, mean, Elizabeth, I'm actually gonna show this on screen, put it well, wow, this is essentially building your own knowledge bank that captures the problems, lessons, and insights from everything you do.

19:54 And that's exactly what this is. So um I am excited to show you how to do it.

19:59 A quick question, we're gonna actually get to it, but um somebody had asked, um I can't find, there we go.

20:08 Craig, how much time do you, on average, do you find you're spending on maintaining these spreadsheets? Hmm.

20:15 So maintaining, I'll put it this way. So after a client call, I have a checklist that I go through.

20:22 And so I'm downloading the transcript, I'm putting it in a certain folder, I'm running the prompts against it, and then uploading the output from those prompts into these spreadsheets.

20:31 And that take, I'm going to do it with you here in a minute and that takes minutes. So then the time that you spend going through and kind of pouring through the data, you can spend as much or as little time as you like.

20:43 In the different seasons of my business, I spend more or less time doing it depending on what I need.

20:50 But in terms of the mechanics, it's super fast.

20:53 And what I'd like to do, I haven't figured exactly how to do it yet, but like create an automated workflow to where I'm dumping the transcript into a folder and it kicks off a

21:05 process where it runs the different prompts and uploads the sheets for me and I'm just time constrained. I haven't figured it out all yet. always like that.

21:13 You're like, oh, could make this all automated. But you're like, but I don't have the time to end up doing this, which will eventually save you time. But like, who has the time for that?

21:21 You will. Cool. Let's dive into what you're doing.

21:30 So I just want to start off the top and say, what Craig is initially doing. is he's recording his client calls on Riverside, because we've gotten that question before.

21:40 Recording it on Riverside and getting that AI transcript. And then tell me what you do with the AI transcript from there.

21:48 So I'll uh download the transcript from Riverside. And I just put that on my machine.

21:55 And then once I have that, I take it and I upload it to an AI model and apply a prompt to the transcript.

22:03 And so the output of that prompt basically is a comma separated value file that I save and then import into Google Sheets.

22:13 And so do you write off the bat I'm just gonna go step by step. You download the transcript, are you doing SRT, TXT, does it matter?

22:21 Okay, TXT file, you take your TXT file, you download it. And then which model do you like to use best?

22:30 So I'm using Perplexity right now. And what that does is that has basically a model router. So it looks at the requests that you sent.

22:39 And it figures out which model to use best. So generally, I like uh Claude37 for writing.

22:46 um And I like the sort of whatever router model that Perplexity is going to pick for me in order to do the analysis and the post-processing.

22:57 Everything that I'll show you works on ChatGPT. It works with Claude It works with any of the models that you have available to you and they also work with the free Versions of them.

23:08 I'm using my paid Perplexity version just because I've been iterating on this and I need a lot of requests to make sure it was set up right ah But I mean really most of the models

23:18 that are available readily available to us right now and even available for free Do a phenomenal job of doing this type of processing Awesome.

23:26 And if you are using a free model, like ChatGPT's free model, I do it. We all do it. So we will accept that you.

23:35 So you will find that if you're having a longer call, your transcript's probably going to be way too long to just copy and paste into ChatGPT.

23:43 So Craig has included an incredible link that I'm going to put in the chat, and I'll also send to you all later. It's a text chunker.

23:51 So you can take your entire transcript in that TXT file.

23:56 and put it into this text chunker and it will break it up into sizable pieces that ChatGPT will process so that it doesn't say it's too many characters.

24:05 So really useful tool, bookmark that thing. It is a lifesaver for me.

24:13 I'm glad you found that. I added it in the links there. Is it? That's awesome.

24:21 Yeah. Well, my background is in web development, but I just generated that tool out of Perplexity because it was phenomenal, the features that it even put in there for me.

24:34 So yeah. you do that. Let's go through the next process.

24:42 Okay, well, so why don't I let me show you the prompt that I'm using.

24:49 Yeah, let's do like an example of it and then we can see how it's working. okay, so here is.

24:55 um So I'll show you the prompts and then we'll use it and you can see the output from it.

25:03 And I like to explain it kind of line by line because everything that's there is there for a reason.

25:08 So anytime you're working with a model, generally what you want to do is have it adopt an identity, give it a role.

25:14 So here I'm saying you're expert researcher, you're going to look at my call transcripts. And what I'm having you do, AI model, is create a CSV ready table for me.

25:23 So that's kind of where we're headed with this.

25:26 I want it to analyze the transcript and identify all the client pain points, struggles, frustration, roadblocks, problems, anything that you can see and kind of come up with

25:36 that, uh, my client is as voiced within the call, but skip personal stuff.

25:42 We'll, we'll keep that out of the analysis and then I'll have it go in. It knows who I am, but go ahead and pull the client name out of the transcript so that I can feed it into the spreadsheet pretty easily.

25:54 For each problem, what I wanted to do is essentially this is how I'm asking it to build up the table is kind of title the problem in and of itself, and then give a very detailed context of the description.

26:06 Now, usually this isn't enough, and I'm going to show you how we'll have to push the model again to give us even better description, but this is a good starting point.

26:14 And I say, OK, associate an emotion, give me that need statement, and then suggest a solution.

26:20 uh I'll switch back over to the other screen now and we can kind of see what the output looks like um from the model itself.

26:30 And again, we're going to be sending you these prompts. So you'll be able to essentially just copy and paste this with your own transcript, like send your own transcript, copy and paste these prompts and be able to do exactly this.

26:41 Exactly. Absolutely. So here's that whole prompt that I put in there. oh And all I did was I attached the call transcript to the chat thread.

26:51 So that's there. And so here it is. It's going through and it analyzed everything for me and it gave me this table. And this is good.

26:58 So this person, Edward Awesome, I have clients with names like that, ah he needs to update his course.

27:07 And so he's feeling a little uncertain about it and kind of trying to figure out how to make that happen.

27:12 ah I'm sure about review processes, it picked out everything using his wording, his terminology.

27:19 So once I got that table, was like, okay, that's good. But now again, I want to push the model even a little further.

27:26 So I go back to it and I say, okay, are you sure that you found everything that my client has had an issue with?

27:34 So I tell it to go back and say, carefully read the transcript again, and make sure you didn't miss anything. And then once you're done, update the list.

27:42 So here I'm telling it to make sure to add in anything that found in the past. But this is really the key to the second layer of prompting.

27:51 Just elaborate more, give me more detail, because what I want to be able to do is go back to this spreadsheet six months from now, a year later, and be able to understand that problem from the description without having to need a lot of context.

28:04 And so this is the most important part of it. And so then I just kind of do some house cleaning things, only use the provided reference, don't make anything up, please, skip personal problems and don't mention me.

28:17 So now the table that it's generating for me is much more elaborate in terms of the description.

28:23 And so this is a lot closer to what I would want to be able to log within that Google sheet and have available to me.

28:30 Now, within Perplexity, it makes it really easy because I can come down to the bottom of the the table here and I can just click this little icon for download CSV and that'll make it available to me.

28:43 If you're using ChatGPT then you might have to tell it okay format this as a downloadable CSV and in fact you can even tell it create the file for me and ChatGPT will do that for you.

28:54 It takes a minute if you're on the free print plan like two or three minutes but it'll just create the file for you and then what that looks like is something like this.

29:06 So here is uh the CSV created for my problems.

29:13 And so everything that was in that table is now just in this comma separated list. And the only thing you want to make sure to do is to take out the header row.

29:22 And then once you have that, now I can go back over to my spreadsheet. you doing? Are you like copying and pasting this now?

29:31 So that's just a file on, so it created a file on my computer. And so it's just a CSV file.

29:39 And then I can come up here and go File, Import. And then I can upload my problems.

29:47 And then I just want to tell it to append to the current sheet. And then it's separating the values by commas. So I'm just picking that here.

29:56 And then once I import that data, here's Edward Awesome's problems. from the transcript. That's really cool.

30:04 And you could even take it a step further. You can like take that spreadsheet, download it, and then say from all of these problems, like you could say, like, wow, I'm getting excited because then you can take all of those.

30:16 You can take all of the client calls, put them all into the spreadsheet, and then be able to take that entire spreadsheet and say from this spreadsheet, what are the most common problems that people are facing and what are the best solutions for this?

30:29 Like, yeah, there's so much that you can do.

30:31 And then you can take on, like, that could be marketing campaigns for you.

30:37 Like, that could be the next course that you teach, that can be the next email newsletter that you send out, or video that you create, or maybe podcast episode.

30:46 Like, use this as a way to inspire the content that you create, because people are looking for the solutions to their problems, and now you have this document.

30:56 that has all of the problems that your clients face and you know how to kind of cater to those issues. that is super cool.

31:04 And Travis, I'm also gonna laugh a little bit because I also thought this when he said I want to import all my problems. I really felt this comment.

31:13 So yes, agreed. I wish I could upload mine and get that fixed. Yeah, that would make the process so much better.

31:23 And you even read my mind a little bit because one of the other demos is to show you how to take all that data in aggregate and then basically look at your clients from as a group or as individuals.

31:33 So that's really cool. I know. All right.

31:38 So instead of just going back to the slides, so let me show you the next one just in here.

31:46 Basically what we're doing is we're looking to extract the lessons at this point. Process is exactly the same.

31:51 So what I'm doing is I'm saying, okay, look at this transcript and then look for all of my lessons and the strategies, tactics and insights only using my words as a reference.

32:01 And one of the things that Perplexity can do when you submit a prompt is you can go in and I'll come back and show you how to do it.

32:09 But you can turn off web searching basically. So it's only using the uploaded file as a source, which is really nice. Why would you want to do that?

32:17 em when I'm trying to capture my lessons, I want to know what's specific to, to me and my custom IP, my, my custom vocabulary that I use with my clients.

32:29 I don't want to try and catalog other people's ideas. Certainly I'm influenced by other people, but, um, I just want to figure out what am I teaching specifically here?

32:39 Because think of it this way. You could use this, this lessons library that we're building. And let's say you want to write a book.

32:46 or you building a podcast series or anything that's specific to you and your course or your coaching business, I only wanna know about what's my information.

32:57 So that's why I constrain it that way. And then just go in for each lesson, give me a one word category, title it under 75 characters, detailed summary all along the way.

33:09 And so here though, I'm asking the model to compete with itself.

33:14 So at this point, say generate at least three drafts and then go and look and select the best option based off of those three drafts that the model is generating even before it

33:26 returns a response to me and pick which one it thinks is best based off of how it'll best serve my client.

33:34 that's kind of uh just giving it a chance to reason a little more before it's giving you a response. And so here it's.

33:43 it's outputting the different drafts. I have a table here is draft one, table here is draft two, and then draft three.

33:50 And it says, okay, well, here's the selection and here's what I picked for you. Okay.

33:55 So now I have this third iteration of, of the, you know, what it found within the transcript.

34:03 And then I go through that same process again, in order to say, okay, thank you model. You did a really good job, but at the same time, I'm going to push you again.

34:11 And go back and reread the transcripts and make sure that you got everything that I taught out of that.

34:15 And then also go through and elaborate more so that I have, can basically it's not making assumptions.

34:23 I can just go back and read this lesson and have full context for it. And so then it gives me this table here. You can see it's fleshed out a little bit more.

34:31 And now I have a comprehensive list of everything that I taught all within this one coaching.

34:40 That's amazing. That is very, so let's talk about, it's the same process then as you did for the last one.

34:46 You would just get that CSV file, put it into your spreadsheet. What could you do with these insights?

34:55 Well, like I said, going back and trying to figure out how you're using what you're teaching has so many different applications.

35:04 Like oftentimes I think in terms of, I have a body of work that's in me and that body of work is expressed in a number of different ways.

35:12 Sometimes it's just me talking to a client, a coaching call. Sometimes it's me being interviewed on a podcast. Sometimes it's a blog post that I'm gonna write.

35:21 Maybe it's a YouTube video that I want to create Maybe I literally just want to write the book based off of my body of work This gives me a lion's share of what I would need in order to make that happen.

35:36 It's not written yet It's it's not developed yet, but all of the ideas are there and so uh It just gives you the documented foundation of your body of work

35:47 And so I think if we wanted to move on to writing in your voice, do you think we can, is that where you're gonna go next?

35:59 Okay, cool.

35:59 Because I think that what we talking before, there is like, when people think about writing for you, like AI writing for you, I mean, I'm gonna admit, I'm a former content writer.

36:10 And so I can spot a ChatGPT written content usually right away. right. Right.

36:18 so I feel like justice for the M dash, mean, like everyone, it's like, it's using the M dash all the time and emojis and always suggesting that you wanna take it to the next level, which not justice for that term.

36:33 And so I wanna talk about, think that that is how a lot of people view having AI right content for you, especially based on.

36:42 some of the stuff that we saw in the beginning in that word cloud. So let's talk about how you can really take your transcripts and have this, an AI model write more in your voice.

36:54 So I have a son who's in college right now and he is terrified that he'll use a model to help develop his thoughts and build outlines and stuff like that.

37:06 He's just like, if I get dinged for using an AI model on any of my stuff in any way, he's just paranoid that that's gonna happen because obviously he doesn't let the model write for him.

37:17 So I told him I was doing this and he was like, all I'm gonna test you out. And so he took one of the emails that I wrote using this process.

37:23 and the evaluation software came back said it was written by a human. And there's a specific reason for that.

37:32 And that's because I'm not asking the AI model to write for me kind of out of pocket.

37:38 What I'm doing is I'm saying, okay, number one, here's all the context of everything that I wanna say and I want to teach in what I'm writing.

37:46 And number two, here's a series of examples of my writing.

37:50 And so you're going to read and analyze my examples and then you're going to take my thoughts and my words and then you are going to write on my behalf, but as adopting me and my style.

38:00 And so this is how we move out of the cringy. This is how we move out of the obvious AI written content because it truly is a hybrid between what you're writing and having the model help you.

38:12 So let me. oh, I was gonna say, so this kind of starts with, because we had talked before, this starts with an examples file and a context file.

38:22 Let's first talk about, because that's kind of before you start using this prompt, what are those two things?

38:28 Yes, so I have my context text file right on the screen right now.

38:33 Now this was me dictating into my phone this morning a recollection of a client call that I had and a problem that I was sorting through with one of my clients.

38:45 And so I just kind of, you know, for about five minutes, four and a half minutes, I just kind of put out my ideas from there.

38:55 And so this is just a free form context of what I want to talk about and what I want to have right about.

39:00 So this is the context in which I'm giving the model say, okay, this is the topic you're going to talk about.

39:06 I'm not expecting you to go out to your trained data and use probabilities to figure out the next word necessarily.

39:14 This is what you're going to write about. Okay, so that's the purpose of context. And then I come into examples and here are five different emails that I've written by hand.

39:24 using my tone, my voice, my approach, whatever, my formatting, all of that.

39:29 And so I don't know if you're familiar with like HTML or tags or anything like that, but I just basically wrap each one of them in an example tag.

39:38 And then at the bottom, you'll see I close it off with a slash example.

39:46 So it's telling the model within this container, this is a single example. And so I have each one here wrapped within each one of those tags.

39:54 So now when I go in to use this, I can come in to the model and I give it context and examples and ask it to write for me.

40:06 OK. So let me go over to the slides really quick to kind of show you the prompt because I think it's a little easier to digest that way.

40:12 um And then we can kind of look at what it's doing because there's some really specific things that I'm doing in that area to um get it set up.

40:21 So just let me find where that slide starts.

40:26 I like that Candace just posted on the chat her banned list of things that they refuse to, I assume, refuse to put in their content.

40:36 I also have my own banned list that I haven't actually written down, but I really should because there are so many terms that as soon as I see it on an email that's written for me, I'm like, no, I'll write it myself.

40:49 I'll write it myself. Sorry, hold on one second.

40:57 And as he's kind of loading that right now, anyone in the chat, let us know how this is going, if this is helpful for you all, if this is something that you want to apply to your own work.

41:07 Because for me, I think that this is, I want to take all my webinar transcripts and be able to create takeaways from the community.

41:15 Or even I can do this with the chat from our webinars.

41:18 uh almost there.

41:25 PowerPoint is not being my friend right now.

41:29 This is the beauty of live Yeah. And somebody had asked, I put Perplexity, that's what Craig has been using.

41:40 I put the link for Perplexity in the chat, but I'll also be sending it at the end of the session.

41:47 And it is also linked in the document that we're gonna send you. Perfect. just, Perplexity was one of the first tools to include searching before ChatGTP did.

41:57 And so I started using it, but they have a lot of other really compelling features that are available. And so that's just what I've uh landed on.

42:05 ChatGTP, the latest versions of like the website and the experience has even more incredible stuff.

42:11 So it's like, they're always competing against each other for different features. So.

42:16 Sometimes I'll even ask like a few different places for the same question and kind of see which one is giving me the content that I want in that day, you know, so.

42:26 yeah, yeah, definitely. Okay, so writing in our voice, again, the identity I'm asking it to adopt is an expert copywriter for me.

42:35 And then essentially what I'm doing is I'm giving it my ICP, okay? And uh so I'm an online coach, I help course creators who want to scale their businesses.

42:45 This is kind of what I'm about, what I'm doing, and the purpose of why I'm writing. So I tell it I'm attaching two different files to the chat.

42:53 So examples, this is my writing style. Use this to emulate me. Context, this is the content that you want to use in order to try and generate what we're building here.

43:04 And then the instructions are, use examples to learn about me. Reason about the data that's in context.

43:12 And then find three different topics to teach about for the email that I want to generate.

43:17 And when appropriate, go ahead and build a framework for me. So I might use this.

43:21 I probably won't, but it's just nice to give the model a chance to sort of suggest something that I can make more memorable or more hooky as I'm building this email.

43:33 Here, I want it to use web searches. So maybe there's a quote, or maybe there's a resource, or something that's relevant to what I'm writing about that I want it to go ahead and use and include within the text.

43:44 So here, I'm giving it express permission to go and find those things for me. And then I tell it to uh generate multiple explanations for each one.

43:52 giving it that chance to build three different versions of what it's doing and then comparing them and pick the one that's most effective specifically for the people that I'm writing for.

44:04 And then pick the best one based off of that. And then I tell it how to structure the email. I say, okay, give me an introduction, teasing the three topics.

44:12 explain each one that you decided to use, make it short and easy to remember with a sentence that summarizes it at the end, and then give me a conclusion with placeholders for call to actions and all that kind of stuff.

44:25 And then give me 10 subject lines that I might use or I might not, but it gives me uh a place to build off of from giving a subject to that email.

44:35 and you're including the subject lines in your example txt file. No, I'm asking you to generate those for me.

44:42 Now I'm saying when you give the examples of previous emails that you've done, are you also showing it your subject line so it gets an idea? Oh.

44:51 I should actually, I should add that to my examples. I just put the text of the emails in there. Try it out. Let's see how it goes. I will.

44:58 That's a good point. So then I tell it again to make sure to give me placeholders for anything specific and then the type of writing style.

45:07 I asked it to model after the examples I give it, but I also tell it, give it basically some direction about how to write.

45:14 And then I wouldn't use profanity into my writing, so I make sure to use that. And then this is the most important part of the prompt, really.

45:23 before writing, ask any necessary clarifying questions and wait for my response.

45:29 This, again, is just moves it down to the next level of being able to say we're going away from just randomly generating the next most logical word that it's chaining together to what am I trying to say?

45:45 What content am I trying to build? And so by having it stop before it does anything, read everything I've given it and give it.

45:52 giving the model a chance to say, well, what do want to do with X, and Z?

45:56 It makes all the difference in the So now I'll go back over to the browser and you can kind of see what it gave me.

46:06 So here's the prompt as I demonstrated. And then it came back with these questions. So what kind of call to action do I want?

46:15 Should I reference any client stories? do you want me to use it? Sometimes the questions are better than others. These aren't super great.

46:22 But I said, OK, the call to action is going to be booking a coaching call. Just use a placeholder and make sure to use my name.

46:28 It's like, OK, well, here's the subject lines that it generated for me. Might be good, might be bad, but they're a good starting point.

46:35 And then it generates the text for me. So here's the text for the email. And I wouldn't start off with my friend, I would change a few things.

46:45 It gave me a quote. I would take easy takeaway out. But generally, this is pretty much like how I might write an email.

46:54 So this gets me about 80 % of the way to an email that I would write.

46:58 And then I can take this and I can refine it and make it available um to my list based off of a voice note that I created of something that I wanted to talk about.

47:11 That's so cool. So you could just be like walking down the street, record something, and then ultimately it can end up being your email without you really having to put that behind.

47:20 Or take the transcripts from a coach from, from a Riverside conversation that you've had, you know, maybe you find an excerpt of something you want to talk about, use that dumping into context and your writing based off.

47:33 I mean, I'm even thinking about when we had our first call, like if I had recorded that and then just taken the transcript and been like, hey, we're doing a community workshop, it's the first one.

47:41 ah Here is Craig that is going to be leading it talking about his process. Like it could have written that email for me and saved me a ton of time.

47:52 So.

47:52 um if you had examples built up and you know, five's not even a lot.

47:57 You could probably give it 10 or 15 or whatever, just to add more variation and more depth and texture to how you write and have those available as the examples.

48:06 And you can continue to train it as you go too. So you could even say, I've done this in the past. I've sent the final email. I'll say, this is what I ended up sending.

48:13 Please use this. Please use this as context for, or an example for uh my next email that I ask you to write.

48:22 Candice also pointed out, which I thought was funny, I tell it, do not agree with me. Challenge my approach. Otherwise it wants to be your best friend.

48:31 John also agrees with that. That is a good approach.

48:34 was laughing when I was reading the prompts because like it's so I'm so used to my very like kind American-ness of being like, please can you try to do this maybe and it's like do not write in my voice.

48:48 It's like do not write my name. And so it's just like I mean I was doing a AI webinar a while ago.

48:56 And I kept writing please and somebody in the chat was like, do you need to say please? And I'm like, I don't actually.

49:01 This doesn't has no better outcome because I said please, I'm just being way too nice to this robot. Cool.

49:09 So we, I wanted to keep the last few minutes for some questions. Have you all kind of jump in if you have some questions, unless I missed anything that you were going to present, Craig.

49:18 The only one more thing I want to and I can do this really quick is that uh so with notebook AI, you have the opportunity to go in and add a number of like 50 different sources available to you.

49:31 um So basically 50 uh coaching call transcripts.

49:36 And when you do that and put them together in aggregate, now I can come in and I can say, based off all these conversations I've had, what's a 52 week email sequence that I might

49:45 build again, specifically based off of the conversations I've had, my content, my clients problems, what does that look like?

49:53 You can do things like that once you have these transcripts available. So I have one for all of my clients and then I have uh spaces that I create specifically for just individual clients.

50:05 I have this one client where we were working on pricing, she had a bunch of different pricing uh options available over the course of our time together.

50:12 So I was just like, build me a table of the evolution of the pricing of her offer.

50:17 And so I was had that available to me at a moment's notice by being able to know, you know, where we're going to go next.

50:24 so by, ChatGPT uh, the paid version has projects Perplexity has spaces and NotebookLM uh, is the core features being able to use multiple sources.

50:37 Each of these different, uh, resources give the ability to work with multiple sources at a time that persist.

50:44 It's not just for that one chat session, you can go back and continue to ask questions of it. Wait, so back it up because this is very cool.

50:52 So you go into Notebook LM and you just on that sidebar with the plus button, that's where you're adding, what are you adding there?

51:01 The Google spreadsheets? wow. you could, but I'm just taking what comes out of Riverside and I'm putting it directly in there and do anything else with it.

51:11 and then how are you getting it to write this 52 weeks? Can you show an example maybe?

51:15 so with I mean, I would just would come in and type the prompts, you know, I would start off by saying you're an expert online marketer, copywriter, and then I would give my positioning statement.

51:30 I'm a business coach for online course creators do that type of thing. My ICP is these people.

51:37 So here's who I'm trying to reach. Examine all of my coaching calls.

51:42 and generate for me a 52 week email sequence based off of the lessons that I've taught and the problems that my clients have faced.

51:50 So when you save the output in NotebookLM, it doesn't keep the prompt. keeps only the output, but that's essentially the prompt that I use in order to generate this.

51:59 And this is the outline for 52 week sequence that I could build based off of my specific coaching calls.

52:07 So then you take, no, go ahead.

52:10 So I was just going to say the difference here though is that these outputs, correct me if I'm wrong, are taking it based on all of the transcripts that you uploaded there rather than maybe just one at the time.

52:20 Super cool. So you could do that for all your clients or you could do that just with one client.

52:24 So if I need to remember something specific that I, know, ongoing conversation that I've had with a specific client, I can do that with these types of tools.

52:35 And then the bonus prompts I have for you take it even a step further.

52:39 So after the, the, calls over, you can apply this prompt to it and it says, okay, well, where, where do the, Where do we leave off?

52:47 What are the points that we need to continue going on? And what are some kind of novel and sort of probing questions that I can ask my client the next time around?

52:57 And so preparing for that next meeting with them becomes even a lot more easy by using these types of approaches.

53:05 Very, very cool. I feel really inspired by everything that you're really sharing here. We do have a few questions that I want to get to.

53:14 So one that came up a few times was, can Craig talk a bit about how he tells clients that he'll record their conversations and what he'll do with the recordings?

53:23 We did get a few questions about revealing to your clients that you are recording and how to make that a comfortable process for everyone.

53:30 For me and for my coaching business call recordings are a feature of the the engagement and so I tell all my clients I'm recording all these calls and in fact after each one of

53:41 the the calls I maintain a document where I put a link to the call and this is another part of the bonus prompts that are available in the link with the the notes that we have for you that It's gonna look and basically figure out the client actions.

53:55 So it's gonna read the transcript and say okay.

53:57 Here's the stuff we talked about Here's what you need to do So for my business, it's just kind of ingrained into how I work.

54:04 So there's no surprise. And of course, anytime, if we talked about something that shouldn't be recorded or processed or anything like that, my client would tell me and I would not do that.

54:14 But that's never come up for me.

54:17 Another question from Damian, uh Craig, do you need to give it all the reference material every time that you go to write something new, or will the AI retain the original learning?

54:27 So when I showed you how to write an email, basically I'm just giving it the context, which is basically the information that I want to talk about within that email.

54:36 And then I'm giving it the examples and I need to do that each time. So I'm creating a new chat thread. I'm uploading those files and then I give it the prompt asking it to write in the way that I told it to.

54:46 So in that way, ah you need to do that new each time. But once you create that thread, if you go back to it, you can continue to refine the text that it generates for you.

54:56 If you use Notebook LM or you use Perpeccly Spaces or ChatGTP projects, you can upload those files and those stay there.

55:04 That persists that context for you. So you can continue to go back and create new chat threads based off of what you've em made available to the model there.

55:14 So it's a little different depending on what you're trying to accomplish. We have another question from Adrian.

55:21 You had talked about the copy AI checker. Which one was that? Do you recommend one? um I don't know off the top of my head which one he used.

55:30 I know Grammarly has one. If you search for it, there's a number of them readily available. Awesome.

55:37 then um Amy had said another about what you're using. Your slide deck is awesome with animated follow along graphics.

55:45 Which platform do you use to make that? ah That's PowerPoint and the key there. Well, the key there is the morph transition, which basically does tweening for you.

55:55 So if you get good at just basically essentially building keyframes, where something starts here and it ends there, if you use the morph transition, it'll do the animation for you.

56:05 There were a few questions about protecting intellectual property, protecting the privacy and intellectual property or client patent patient content.

56:19 Can you speak to that on how you kind of protect your clients while doing this process? Not a lawyer. So I can't speak to the specifics.

56:29 Basically, each one of my clients knows that this is a part of the process of our working together.

56:35 And we don't discuss sensitive financial information, and we don't discuss, you know, deep personal issues.

56:43 Like, if you're a coach, and you're a mental health coach or something like that, this process probably is not for you.

56:50 But in a business context where I'm just help more But I'm helping people scale their course businesses.

56:57 The data that I'm working with isn't sensitive enough that I have to worry about protecting it in that way. But I do have consent from my clients in order to be able to process it like this.

57:09 Awesome. we have those were basically all of the questions. I mean, we had a few more, but there was a lot of overlap in some of the content that people were asking about.

57:18 And so I just want to take a minute to show you a few resources before we sign off. First, if you're watching on Riverside, this should be a clickable.

57:27 link, but you can find Craig's prompts here. I'm also going to put them in the chat and you'll get them in your emails for everyone who registered for the session.

57:40 And I also wanted to point you to a bunch of free downloadable guides.

57:46 that we have in Riverside um that are created by creators, for creators on different parts of the process, including AI.

57:55 So if you want to check out any more resources from the community here, let me know. Check out these guides. And I am always open to even more ideas.

58:05 And that was our first workshop. So this, as I said, is a new program that we're doing, a new series that we're working on.

58:15 So I would really love to hear from you how you liked it, what you found valuable, what you'd like to see in the future.

58:21 And even if you want to host your own of these workshops, we're looking for people to share their processes with others so that we can all really learn and work together.

58:30 And I think that I speak for everyone, Craig, when I say this has been really interesting, valuable, inspiring information.

58:42 And so I am really glad that you were. here with me to host this very first workshop because yeah, of course.

58:51 And until next time, Craig, you're gonna stay on the call for a second, but until rest time, I'm kicking the rest of you out of this workshop.

58:59 But thank you all so much for joining and I can't wait to do this next one. This was a lot of fun. So until next time, talk to you soon.

59:07 Thank you.