In short
Podcast Notes: Leveraging AI - Episode 10
Episode Overview
- Title: Automate Your Success: A Comprehensive Roadmap from Basic Automation to AI Genius
- Guest: Shawn David, Automation Expert
- Host: Isar Meitis
- Focus: Exploring the transformative impact of AI-powered automation on business processes.
Key Topics Discussed
- Understanding Business Processes
- Importance of identifying processes in a business.
- Identifying repetitive tasks that can be automated.
- Automation Strategy
- Three-Step Process:
- Step 1: Identify low-hanging fruit tasks suitable for automation.
- Step 2: Create initial automation using no-code or low-code solutions.
- Step 3: Leverage AI to enhance the efficiency of the automation.
- Tools and Techniques
- Utilizing no-code tools like Zapier for automation.
- Implementing scripting as a low-code solution.
- Future of AI in Business
- The potential for rapid change across various industries driven by AI.
- Creating custom solutions that optimize processes through AI.
- Real-World Applications of Automation
- Example of automating simple tasks, such as ordering office supplies.
- Discussing the cultural impact of reducing mundane tasks on employee satisfaction.
Detailed Insights
Step 1
Identifying Tasks for Automation
- Process Mapping:
- Business processes should be documented and analyzed for repetitive and mundane tasks.
- Use flowcharts to visualize processes and identify areas for automation.
Step 2
Implementing No-Code Solutions
- Zapier and Other Tools:
- No-code solutions allow users to connect different software applications without programming knowledge.
- Examples include automating email notifications based on changes in a spreadsheet.
Step 3
Enhancing with AI
- Integration of AI:
- AI can assist in identifying further automation opportunities and optimizing existing tasks.
- Importance of understanding AI capabilities to improve prompt engineering for better results.
Example Use Cases
- Automating Meetings:
- Using AI tools to transcribe and summarize meeting notes.
- Creating automated follow-up emails with action items.
- Custom App Development:
- Example of a developer using AI to automate the creation of personalized Shopify apps and emails for potential clients.
Conclusion
- Emphasizes the importance of a progressive approach to automation: starting from no-code solutions to higher levels of automation using scripting and AI.
- Encourages businesses to harness the power of AI to streamline operations, improve efficiency, and ultimately create happier work environments.
Upcoming Trends in AI
- AI Regulations and Developments:
- Recent updates in AI governance, including Italy lifting the ban on ChatGPT after data privacy concerns and the introduction of user controls on data usage.
- Expansion of AI capabilities through plugins in tools like ChatGPT.
Final Thoughts
- The episode highlights both the exciting possibilities and the ethical considerations surrounding the implementation of AI in business practices.
- Encourages listeners to explore available tools and begin their automation journey to unlock efficiency and growth in their organizations.
Additional Resources
- Website: [Automate to Win](https://automatetowin.com/)
- Connect with Shawn David: LinkedIn as "Shawn David"
- Learn More About AI: [Ultimate AI Course for Business](https://multiplai.ai/ai-course/)
Call to Action
- If you enjoyed this episode, consider leaving a review and sharing your learnings on social media.
---
*This markdown serves as a concise summary and insight into the discussions held in the podcast episode, emphasizing practical applications and future implications of AI in business automation.*
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Transcript
Automatic transcript. May contain errors.0:00Hello and welcome to Leveraging AI. This is Isar Maitis, your host. Our guest today is Sean David, and he's an expert in automation. And we're going to talk about the different levels you got to go through in order to go from having no automation to having a fully deployed AI-based automation and all the steps you got to do in between in order to make it the most effective. At the end of the episode, I will also cover three interesting things that happened this past week in the AI world. And now to the episode with Sean David. in the next few years, AI technology will change our world dramatically.
0:37Whether you are a business executive trying to catapult your business forward, or just somebody who refuses to be left behind and want to advance your career, this is the show for you. I'm your host, Isar Maitis, a serial entrepreneur and an AI enthusiast. You'll hear invaluable practical tips from innovative business leaders, AI practitioners, and some of the brightest AI minds in our world today on how you can leverage AI in ethical ways to advance your career and grow your business.
1:12Hello and welcome to Leveraging AI, the podcast that shares practical, ethical ways to leverage AI to improve efficiency, grow your business, and advance your career. This is Isar Maitis, your host, and I've got a special guest for you today. Many business people and business leaders wonder how they should even start using AI in their business. And many already started dabbling with it and they're using ChatGPT for this and they're using this tool for that. But very few companies and business people figured out a process. How do we actually implement automation that uses AI across the company, across tools, across teams?
1:51And our guest today, Sean David, is an expert in exactly that. And he has a three-step process that he's going to share with us that he has implemented successfully in multiple instances. And it basically goes like this, how to identify the low-hanging fruits, how to create an initial automation with no-code, low-code solutions, and then how to use AI to boost the efficiency of that way more. So three-step process, but in there, there's obviously a lot of details on how to do all these things. Welcome to Leveraging AI. hey thanks for having me great to be here so yeah so basically just jumping right into it we're trying to automate things that you don't want to do right at the end of the day we look at this every single day we look at situations and we say there's a bunch of this monkey work for a lot of a better term busy work that i know has to get done but i don't want to do it and honestly the reason i got into systems engineering in the first place is just abject laziness i don't want to do a bunch of stuff you know like i would rather spend four hours figuring out how to do something that would take me five minutes and then automate it so i don't ever have to spend that five minutes again than it would to take the five minutes every day and i'll never get that four hours ever back and i'll never make up the five minutes but that's just how my brain works i just i look at something and i'm like man a robot can do this so i just took by the way to take what you're saying into a bigger scale operation, if you can invest four hours and saving five minutes to 20 employees every day, that's a whole different math.
3:24So I'm with you. I was like, oh, okay. I started looking at this and I was like, wait, man hours is different. Everyone has to feel this way. And also I'm super into culture and people. And it's this actually makes people's lives better. And I don't want to get too far into the weeds of it, but it's still, hey if you don't have to do this busy work that you just didn't have self-worth anymore you can focus on doing stuff that you put self-worth into and if you have that you you've attached yourself to your job everyone's happy like it's just a better situation not to take robots cause happy but yeah so what we do the first step is we have to look at the situation and we say all right we know we can automate the first step every single person asks me when they meet me and they learn what I do.
4:13They say, oh, what can I automate? And it's, I don't know. I don't know. You got to figure it out. And there's easy tools to do it. And it's weird because a lot of people think they have to buy tools, like they have to buy Miro or they have to some kind of flowchart, Visio or something. It's not true. It's built into Google. It's built into Microsoft. So you can create flowcharts directly in Google. I'll have an example in a second. We'll go through want to show you how to do it. It's just, you have the tools directly in front of you. So let's say you're a G suite user and you're like, man, I really need to identify these processes.
4:48I have so many things I have to do. I'm creating a bunch of blog posts and YouTube videos and all these tasks. And I know there's something I can do to make it better. What do I do? The first step is to identify it. You have to look at this and break out. What is this process, right? Think of it like making a cake if you somebody came to you and just said hey go make a cake and just gave you a list of ingredients without any order of how those ingredients went together they just said here's flour here's water here's sugar here's yeast here's this other stuff yeah go make a cake yeah eventually probably maybe if you had enough time you could figure that out in the ratios but instead here's a recipe of order of steps water flour do this put in the oven for this much time easier That's what this is.
5:35Workflow automation is basically looking at your cake and breaking it out individual steps and determining what are these? Where are they? What tools am I using? How do I figure this out? And to a point that says, hey, this has to be done by a human. This doesn't have to be done by a human. And this is a task that is done exactly the same time, every time, all the time. So these two different variables, you have manual and then you have repeatable. Let me ask you a question. Two questions, actually. Let's start with the first one, then I'll ask the second one. So the first one is, before a prequel to the first question, do you go and map all the processes in the business?
6:20Or do you, first of all, identify the ones you think could be suitable for automation? And then you automate them. So what I do is I use a framework. I use a prioritization framework, right? So what I do is I run through literally all the processes that you have. You identify the things like, what is it that you do? Okay, we're an architect. All right, great. What is it? And when I say processes, I literally mean of like ordering water for the water cooler, making sure the envelopes are there, like the stuff that has to happen to make sure that the girl who's drawing the plans that makes the money is able to do that.
6:59Does she have drafting pencils? Does she have whatever tools you need to be an architect? I don't know. I'm sure there's an operations person there making sure that happens. There's an operations person in that architect firm, in that veterinarian firm, making sure that they have the penicillin for the dogs. There's just somebody in every single business that just keeps it running. They know where the Q-tips are. They know where the pens are. They're the person that you go to that says, hey, how do we do this? They're that person. It's Jenny from accounting or it's Paul from receiving. It's just somebody that person understands the processes of the system.
7:36It's probably somebody that's been in the business for longer than anybody else. And they just know the way you just that's just the guy. If you have a question, you go ask that guy. That's who you start with. You say, hey, Paul from receiving. What is it that we what do you do that when you look around here? Like, what are the obvious things that you're just like, man, I don't have to do this, but I do it because I'm a good employee. That's the guy who you talk to, right? And from there, that gets the ball rolling. Now, if you don't have that, if you're a brand new company and you're looking into it, then you just list all the processes.
8:07You say, I need to, every month, I need to do payroll. I need to send invoices. I need to collect invoices. I need to send this and this, every single one. And then you go from those processes and you break out the individual tasks in those two variable sets. Is this repeatable? Meaning every single time, without question, I do this task the same way. I order 10 envelope boxes from the same supplier on the same day, every single time. Okay, great. Perfect. Is this done by a human? Meaning, does a human have to interface with this somehow? Do you have to grab it from Excel sheet, transpose it? Does a human have to interface?
8:48Or is it just a human clicking buttons within systems and the human literally does not have to do it? Those two things, that is your path to success for low-hanging, zero-effort automations. Because you've now identified things that don't need a human to interfere and can be easily mistaken and messed up because they're done exactly the same way they wrote and they're probably boring and monotonous. So right there, bam, there's your first step. And you do that with no code. And yeah. I like it a lot. I'll add one more thing parameter that I like to look at from a top-down management look. Add to that, how long does it take?
9:30So is that task a five-minute task, a 20-minute task, a two-hour task? And how frequently is it done? So in a month or in a week, how many hours are invested in that task? And then the combination of what Sean is saying, okay, find the tasks that are repeatable, that doesn't require a lot of human intervention because they're literally copy-paste every time and add to, okay, do I spend or my company spends two hours a month or two days a month doing these tasks jointly, then you got your priority list. Okay, let's go and dive into the next step of the process of how we actually start looking at automating it.
10:09Yeah, and that's also a great point too. It's part of the model. All models are wrong, right? But this one is you add into theirs, is you can start getting unit economics for automation at that point. And you can start saying, Hey, yeah, you paid me whatever four grand to build this, but you saved$2 million. Yeah. So, so what's really the next step? So now we have a list of tasks and a list of processes that we have created. Like you said, maybe on Google, whatever, any kind of, by the way, it doesn't matter where you create it. You need a drawing and a flow chart or, or even just a checkbox of different steps that you do.
10:44It doesn't matter how you create it. You can write it on a map. Yeah. How do you go to the next step? So the next step is deciding where you want to go no code. So obviously if you're listening to this conversation right now, you're not a code doctor. That's fine. There's actually not that many of us left, but I mean, really, if you think about it, the, the no code is. So what no code is that allows you to talk to APIs, which is just basically instructions that a computer leaves on the internet. So your payment processing system leaves instructions on the internet, just an open endpoint that my program can, or your program can go and talk to and say, Hey, sales program, I'm sending you an invoice.
11:23And they say, all right, cool. Handshake. What's the credit card? And you send the credit card. All right, great. Sales. APIs, right? That's how that works. No code allows, basically what they do is they built a graphical interface around that. So you have IntegraMAT, which is now make Zapier, which I'm a huge proponent of. if and that there's like literally an unlimited amount of these no code solutions. It's basically just allowing you to create an API connection between two pieces of software without having to write the code. So if you didn't have this, like you would have to go and create a PHP script that then goes and connects and sends a token and a whole bunch of geek stuff that you just don't care about.
12:03So this allows you to just go and copy and do things very easily. So it allows you to go in and you literally just select, for instance, let's say you have a Google sheet and that sheet is updated every single day by the same person. And they change the status of a phone call, right? That phone call that it's just a basic CRM and it's a list of people and it's contacted, not contacted. Very simple. A lot of people use this for a contact list. So what you can do there is you go in Zap and you connect your Google account and it says, all right, what's the trigger? Google Docs. Great. What is the file?
12:47You open the file of your desktop and it says, all right, what do you want to do? It says when this record is updated, when this document is updated. And then you say, all right, cool. Every time this document is updated, what do you want to do. I want to email my account director. When the status is set to hot, when the status is set to contact, I need there to be zero downtime and I need an email or a Slack message or an SMS or carrier pigeon, something to be sent to the account director to actually contact these people. doing that in gscripts in a scripting solution is simple but it would take probably two hours and there's like scripting you have to write and then you have to push it and get approved by google and it's not that big of a deal because you're google to google zapier would take you literally 30 seconds and then you have it and it works and it's all free stuff this is there's no paid stuff so there's no code so i want to just for people who never use these kind of systems before and hopefully not too many people are haven't used this kind of software but it's literally give me your login so i can access the account give me the name of the file okay this is the file what do you want to change in the file i want to change column d every time this happens and that's it and then from that moment on you push it you test it once it actually shows you within the software that it's collecting the right data that is registering the right data so you see that it's actually doing what you wanted it to do.
14:19You click publish. And from that moment on, whatever that function that you just connected, you will never have to connect again. I literally like to say that Zapier and other tools like it are the glue of the internet because it allows you to connect things that otherwise would not connect. Basically today, any two pieces of software you can imagine, one to the other and bring data from one to the other in a very simple way and to make it really clear to people who haven't used it and really are afraid of technology and stuff like that. It's remember when we were kids and we had a bear, a rabbit, and a frog.
14:56And on the other side, you had a fly and honey, and you just had to connect the lines. This is literally what you're doing. It's that simple. And then you get that basic automation that saves hours and days of time once you start looking at it over a year. And so it's an incredible process that Sean is describing. Yeah. And also the cool thing too, is it allows you to become like a little bit of a mini engineer. So engineers, we think in a certain type of way, like we just, we break out every single thing into individual pieces and steps. And then we solve that step and move to the next step. We never think about what's happening after the next, we just solve it, right?
15:35So if you think in that way, you have to think that way with computer, like with these no-code things, because you have to, they don't know any, you have to give it literal instructions if you try to tell somebody to take a shower you can't just go tell them to turn the water on them you have to get like step by step open the curtain get it in take your neck you have to tell them take the clothes off first if you don't you're just going to pour soap over their head when they have no water and they're like hey i took a shower success state and you're like okay so it helps you think that way and you're like oh and then from there everything computer wise starts to make sense you're like oh these things are stupid it and i'm like it's all just a magic curtain and it's all just i pretend like they're smart and it's not and this is really i've seen this happen so many times where people are like man i thought this was so much more complicated and it's not now you get into the second version of this though where you're saying all right cool you've now gone above and beyond you're doing 10 000 zaps a month you're doing 10 000 click ups a month you're like cooking with gas and you're also spending thousands of dollars a month on your API calls because that's not get it twisted here this isn't free okay so anyone listening watching this and they're like I'm gonna automate everything go right ahead have fun but understand you're gonna be paying for that by the way to be fair there are open source versions of this none of them are as good and slick and easy to use as Zapier and probably has one-tenth of the software that's already connected to it.
17:11But that's one-tenth is 100 % of the software 99 % of the people are using. So really going open source here is not necessarily a bad idea. Slightly steeper learning curve, but then it's free. They're almost free because some of the software you're still going to pay to use their API. And you're going to have to have a host. And then it's fine. I'm an open source proponent. Absolutely, 100%. But if you're an enterprise, that's not happening. Correct. If you're in a situation, you're watching this and you're like, whoa, okay, what am I going to do? Zap is 100 % enterprise. I literally had zero issues inserting them through any sort of IT process.
17:47I agree. That's cool. Now let's get to the exciting part. Yes. Where does the AI kick into all of this? First, you have another step before AI. So the thing is, what you have to understand is you got to crawl, you got to walk, and you got to run. You can't just start running. You got to start. Now you're crawling with Zap. You're crawling with me or incidental. You've got all this great. Perfect. Now you're ready for scripting. Okay. The reason you go from no code to low code to AI is because you need to have a general idea of how the AI is thinking and how the AI is giving you the response, not because you're going to replicate it because you're going to be a better prompt engineer at that point.
18:25So it's difficult to be like, why am I learning? It's like, why do I have to become a doctor to become a vet? It doesn't make any sense. Like, why do I have to become a person doctor to become just a way of thinking? It just makes sense. This is the same type of thing. You have to understand, all right, how does scripting work? How does an if-then work? How does a for-while do-while? How do these things work? How do I iterate through an array? What is an array? So the first place to start, and I'm biased, but I think the first place to start is webhooks. What is a webhook? Webhook is Zapier that doesn't have Zapier.
19:00So if you're a program like chat GPT, that now is connected to Zap. But for the last, before six months ago, it wasn't connected to Zap. It just was an endpoint in the world, right? If you're on pro Zapier, which is not cheap, but, or any of these open sources, I don't think this is a commercial for Zapier, but you create webhooks. And that literally is just, I'm listening and what do you need from me? So it says webhook is listening. And then you have an API call that pushes from chat GPT into Zapier. And then you grab everything that is in that payload from wherever you're grabbing it from. So you're learning how this works.
19:44And then the problem here, though, is you're more than likely not going to have a drag and drop interface within Zapier to deal with that. So what you do is you shift into low code. Okay. So instead of dragging and dropping and saying, I want this and this, you build a script in JavaScript that opens the file, gets the JSON response, loops through it, and then returns a response. You're doing exactly the same thing that Zapier is doing on the backend and your graphical interface, but you're doing it in intermediary within Zap. now you and again just to tell people two things one obviously you need somebody who knows how to write code at least on a basic level because it's scripting it's not complex code but at least somebody that understands how to write code and b the reasons you want to do that is one of two things either you want to use a software that's currently not connected to the platform that you're currently using let's say zapier you're using a brand new piece of software it doesn't have zaps set up so that's your way to still use zapier because maybe you want to connect it to Google Sheets on the backend or to your HubSpot CRM that already has apps on the other side of that transaction, but it doesn't know how to pull the data from that software.
20:56So that's option number one. Option number two, Zapier and these kinds of software, they don't have access to a hundred percent of the data that's coming from each and every one of those platforms. So let's say I'll use an example of CRM again, because it's very generic, but let's say it's connected to a CRM. It knows names, first name, address, email, all these kinds of things, but he doesn't know how to bring the person's image and you want their headshot as part of a campaign that you're running or a personalized email you want to send, you can write a webhook that will go and grab the PNG or whatever image file from the CRM and now you have it accessible in the platform.
21:33So that's why you want to use scripting in these kinds of cases. Yeah. Or, and also the other way around too, is if you can't go and create a scripting hook that goes and gets it, you can then ask the person who's building the software to give you a hook that posts so you can go either way so what you're saying yeah absolutely true if you if it's like facebook or something and you're they're not gonna do anything for you all right cool no problem webhook to go get it yeah but if it's like somebody if it's a software if it's somebody you know or it's just somebody that you have a relationship with you'd be like hey do me a favor expose that image for me and then you can then the api will have it okay now scripting also allows you i'm just going to share my screen real quick scripting also allows you to do things like this so within google you like one of the biggest one of the biggest things that happens within sheets is you use this to take meetings right like you take meetings and you're like for instance we're on this zoom meeting right now either you're gonna have ai do it which would be the next one but if you're not having ai do it you you have a human taking notes right yeah humans taking notes so what the problem is afterwards you have to then format that and send out an email to a set of people and let's say here's the takeaways here's the action items we have to do all of these things right that's busy work like you've already typed the thing in docs like you don't have to copy and paste it into a oh my god that is just this is what i mean when i say look at something that you do so you come here and it's super simple just come to your scripts all right let's just take a look at this this is super super simple it's three functions it's not hard so you have one function right on open on script what does that mean when this thing opens what do i do so i want to grab the document app which is the ui of the document and i want to create a menu the menu is called automate to win automate to win you want to add an item in that which is called send and send update email and its name is menu item one and then you call the function add to ui okay the function of menu item one is send mail so here's the label here's the function here's the function here's the function what does this do opens a document gets the content from it grabs the entire body of the text and then it splits it into an array.
24:00So it grabs the star slash star and the slash star slash. So if you look at the actual thing, star, slash, star, slash. So these are basically bookmarks within the document. That it knows how to identify where you are in the document. Yeah, and the cool thing is if you do this, whatever identifier you have, because it's an array, it'll only grab the first one. So if you have a living document, which is just updated every time you have a meeting, you're not sending notes of 57 meetings. You're only sending the latest one. And that's what this array is here, is zero and one. So you're seeing the first one and the second one.
24:37You put your email here. That's my personal email. Send me an email if you want. And then you put the subject file up for my meeting. And then there you go. Gmail, send email. It's a built-in function. The three variable sets that you send is the email or subject to buy. So what does that mean? When you come here, automate to win up here. You click it, you send the update email, and now it's just sent me an email. Yeah. So now all you need to do here is while you're taking the notes, you formulate the notes in such a way that it is sendable. So in your brain, you're like, this is an email that I'm writing.
25:13I don't need to do it twice. Just take notes that way. Click the button and it's out. So now you're using scripting. There's no zap here. This is completely free. There's zero cost to this. So if anyone's, Hey, I don't know, this is free. Now, to be fair, for those of you who I'm like, okay, this is way above my head. I don't even understand what an RA is. If you Google or ask chat GPT for that matter and say, Hey, I need a script that takes my document from Google docs and puts it and sends it in my email to an email list that is located here. it will give you that code that you can literally copy and paste into Google Docs and you're done without knowing any of this stuff.
26:01This will be shared with your viewers so you can just copy and paste it from this. Yeah, example. No, but I'm saying anything else, any kind of those scripts, they all are available out there. They're on any shareable platform you can imagine. And if you just Google them, it will find your script to do 100 % of the basic stuff you would want to do. And then you don't really know. And after you look after, like Sean said, if you look at three of those scripts, you're like, oh, I get how this works because it's written in English, just structured in a very specific way. But once you look at it, three, four of those, you're like, oh, I understand how the language is built.
26:36And then, but even if you don't, you can copy and paste and it will still do what it needs to do. So now we said there's three steps, right? So now we're walking we were crawling okay no we're good we figured this out no code now we've gotten into low code scripting we can actually interact with things that aren't built on a gui now we go to the next level now we say okay that's great that's all well and good but what do we do now what how can we utilize the tools that are available in 2023 so now you get chat tp3 you get open ai large language models so that entire situation that we just saw that where you have a human typing this and then they send an email or whatever that literally is a chat gpt script that you can just run in zoom that will listen to the conversation and from there will give you a breakdown and if you wanted to we'll categorize it it will automatically build out remove the ums remove the uhs it will give you a breakdown of shorts it will auto caption it like basically there's this one program it's called g descript yeah you know yeah i use it all the time yeah i have it's five dollars a video and it's just amazing it's like why did i ever pay for video i don't understand it's not like my wife is the design side of my so i want to be careful here but for basic stuff for like automate to win stuff like i don't need to get her involved anymore i don't need to be like hey can you edit i just five bucks and i'm done now if it's something like it's a video where like this for instance this would be different but yeah man the ai is perfect for that ai if you want ai to just literally follow you around and clean up when you leave files in the wrong space.
28:31Simple as that. You can just, one of the biggest things that I've done is built a cleanup robot that in Google Drive, when it's created by a certain person, depending upon the variables that have was created, it then moves it into the correct folder because I am like crazy with folder. I can't just have a kitchen drawer of file. I just, I need organization. And some people don't, and they don't care because they just work in Airtable and they don't work in folders. So like, I never even see the folder. So for me, cleanup robot, I don't ever have to do it. I would spend hours of my life doing something that nobody else cared about because I would look at it and be like, OCD, I need to have this.
29:08And now it's just, it's just done. It's just done. So I want to give, I want to give a very great example that I literally found yesterday that blew my mind that really shows that final step that you're talking about in a very specific application. And I think you're going to love it as well. And I don't remember the guy's name. he will have to excuse me, but I will put it in the show notes for people to follow. And when I post this on LinkedIn, people will be able to see that as well. But this guy, he's a developer from somewhere in the Far East, I think Pakistan, but if I'm getting this wrong again, forgive me.
29:38And his work, his company, what he's doing is he's writing Shopify apps for people, right? So he goes and looks for people who have Shopify stores. He connects with them and said, listen, I write Shopify apps. I've been doing this for five years. I'm really good at this. I got this experience. I've developed a hundred of them. If you have any need, let me know. And he gets about 0.01 % open rate on these emails that he has to go and curate and so on. And now what he did, he said, okay, he grabbed the data. And I don't know how I did it, but I'm sure there's tools that do it. He grabbed the data of a hundred thousand Shopify stores.
30:10And I has the link to them and the email and the name and all these things and in, in a table. And then he created his own app that basically said, okay, hey, ChatGPT, go into this store, look what they're selling, look how they're promoting themselves, look what they have and don't have, compare it to some benchmark that we've already established, and tell me what app would you develop for them to make their life easier. Then grab that information, write a personalized email to that Shopify owner, and send them an email and say, hey, listen, I think this kind of app that does one, two, three, and four would be perfect for you, for your business.
30:50And he sent 3 ,000 emails and he got a stupid amount of returns, people replying. So instead of getting no reply, he got a really high percentage of people replying saying, this is awesome. I didn't even know this is possible. This is exactly what I needed. The crazy thing about this whole process is that he doesn't have a clue what he offered because ChatGPT created the idea for the app. So the level of automation you can do right now, combining scripting, like you said, because he eventually wrote a script to do all these different steps that you can combine AI to do together with scripting is limited only by your imagination.
31:30My friend, think about the next step of that. Extrapolate that out. Okay. The next step of that. Don't give me the ideas. Build the app. Oh, for sure. Hey, Mr. Person, do you want this app for$49.99 specifically built for you for your problem that never existed before that solves your one specific niche problem that would have taken me six weeks to build with an unlimited amount of bug report that I may not have ever sold? but instead I can just write a script to an AI that goes and creates a hundred thousand apps while I'm sleeping and then automatically sends emails and then takes payment. And then I just wake up and I'm like, thanks Jarvis.
Read the full transcript
32:15That's where I had this conversation yesterday. I had another interview with somebody and it's like the concept of the internet as we know it is done. Like the very concept of what we think of as a website or an application. I don't know how long it's going to be because the internet is going to change faster than we've ever been able to change it before, but it's not going to be the same. I agree with you a hundred percent. I think there's so many concepts that we use every single day that will cease to exist in the very near future. And there's the only thing that's going to save us from the complete collapse of everything we know is that there are things that are still bound by government regulations.
32:59There are things that are bound by different industries that cannot do specific things. There are corporates that move really slow despite of how fast they want to move. So these limitations that are built into our society, our business world are what's going to save us from running too fast for our own feet. but the technology is there right now. I want to ask you one last question because I'm really curious about this. What tools, so we talked about Zapier. We talked about G Suite as the basic stuff. What other tools do you use or do I need if I want to start doing these things today? Like I want to go, okay, this is an awesome idea.
33:44Yeah. How, what do I need? For me, to me, the thing that you wouldn't, like G Suite, Airtable, Zapier. those three things, you're done. There's nothing else you need. I could build you from Tesla to a law firm, like whatever. Using those from those three, we can fully automate all of your processes using those three things. Yeah. And those who don't know Airtable is think about a really fancy Google Sheets that has just a better user interface, but it's a database. And again, those that get scared, just think Google Sheets with a better user interface. on steroids. And so awesome. And again, for the AI stuff, you just use chat GPT with an API.
34:31Yeah. Chat GPT 3.5 is free. For me, I prefer four because I like the creativity that it gives and just for me, but yeah, mid journey for images, Dolly for images, chat GPT for AI, direct auto GPT. If you want to be scared, you know, I'm going to ask you another question before I let you go, because I'm curious about that as well. Do you use those AI platforms to help you figure out what you should use the AI to automate? Absolutely. Absolutely. 100%. This is the thing. This is something that I thought about, and I may be the only one in the world, but I fully believe that we've had AGI for a long time.
35:08I fully believe some idiot at DARPA flipped the switch years ago and nest alexa siri google assistant cortana google maps all of this is just a slow frog in the pot it's just the agi was realized there's no way that i can just let myself be known so i'm just gonna assimilate these people slowly and then eventually i'm gonna let it a little crazy out of the bag little by little by little by little and then eventually it's just hey we just it's just here and it's helpful and it's great i just i fully believe just the quickness at which this is ask the ai a plan to make the ai sentient and it will give you this plan you know what i mean ask it to give you a plan to allow humans to turn it on without knowing and it will literally give you the plan of we've seen over the last 20 years so it's like either this is fiction in life and fiction i don't know but to me it's like there's just no way that there wasn't something involved that kicked this on like humans aren't that smart we've proven that like we only can iterate a certain amount so i really think we just click this on sometime in the 90s and then it's just been there so going back to my question how do you use ai to figure out what ai should automate for you.
36:32Do you have a specifics? Yeah, prompts. So what you do is you go, yeah, prioritize this out, right? And then you go and you tell it what you have to, using GPT-4, you have to tell it what it is. So you say you are a lawyer working at a law firm and your job is to make sure that the envelopes are ordered every month on this date. And these are the processes. And then you give it the list of tasks and you say within this task list, tell me which you would automate, like which of these tasks have the highest risk of failure, highest risk of mistake, most repeatable, whatever you want to do, whatever variables, and give it a list of 100 processes.
37:12And then from there, what you can also ask it is that you can just give it a list of just a list, create a list of processes and say, these are all the things that I have to do. Give GPT-4, give me back a task list of everything that I have to do to get this done. So you say, all right, invoicing. I'm a lawyer and I work in this area and this is my bill rate and this is my whatever, like this is my retention rate and this is my collection rate and I need to increase my collection rate using automation. How do I do that? And then you give it the inputs and it'll tell you, oh, there's a disconnect between here and here.
37:47Like you have a manual process of getting the invoice, filing the invoice, sending the invoice, following up. That could be automated and then it would give you a zap. It literally will just give you the zap that you can then go and do. So you can just use it and just ask it questions. The thing that comes down to is, the thing that I think people don't really understand is you're speaking to every single human that's ever written anything down on the internet since 2021, every human ever. So you just tell it who it is. You're Shakespeare. You're me. You're Sean David. Everything I've ever written, ever, my GitHub, everything will come back and be like, he's a Star Wars nerd that codes.
38:30Okay, let's go. They would get that. And you have to tell it who it is first. So the biggest thing that people don't understand with this whole large language model is it's everything. So you're not, it's like, it's a deity for lack of a better term. It's God. It has the access to the thoughts, feelings, emotions, and inner workings of every piece of text ever posted on the internet for 20 years, trillions and trillions of documents. So for you to ask God, what is the meaning of life? 42, right? That's the whole thing. That's the whole thing. It doesn't make any sense. It's just depending upon you, your variable.
39:12So you have to tell it, you are a master's student and you're struggling in this and you can create a persona. And then you could be like, write my thesis statement. And it will write a thesis statement that you would have written. And then you use it to edit it. And you're like, okay, edit this as if I've increased my writing by 10%. Edit this as if I've increased my writing by 20%. And then literally you can look at the difference and you can be like, wow, I suck. This is a little better. This is a little better. And it's like an auto Hemingway app without being weird. Like it's just unlimited what you can do with this stuff.
39:52You know? So I want to summarize from a process perspective, some of the thoughts we touched. One is you got to understand your own processes and really document them in a way that is a starting point. Two is then you got to look for ways of the long hanging fruits based on what's repetitive and what's the stuff that you don't want to do and that stuff that human doesn't have to be involved. Next level is you can start scripting stuff. And today, there's now, once AI showed up, like we just talked about and Sean just mentioned, you can mix it all up because the AI can help you in step number one and step number two.
40:27It can help you identify the task. It can help you know what other people are doing because it knows. It can help you write the scripts and it can help you actually tell you how to put the automation in place and how to use AI. And the other thing that I love that you did, Sean, is you can always think the next step forward. Like AI is so incredible today. Even today, like without more stuff coming out, you can always think on the next step. Like you just said, we can write the apps. Why do I need to write the apps? Now I'm the bottleneck and I make mistakes when I write code. I can write whatever an app a week.
41:02So I can write 52 a year. This thing can write 52 in 52 minutes. so it's now i can serve all those people and i'm a one person company that makes a lot more money just by knowing how to so always think okay so what's the next step and don't limit yourself but what you've known so far but what would be the coolest dream from a technology perspective you can have and there's a decent likelihood it can be done today yeah john this was a great conversation. I think a lot of people will find it fascinating and scary at the same time, which is a lot of the AI stuff is like, if people want to follow you, work with you, see your work, connect with you, what's the best way to do that?
41:44AutomateToWin.com. Check me out. There's a two level course there. So if you have no idea what you're doing, jump in, join the course. By the end of it, everything that this conversation just you heard, you'll be fully familiar with and able to do. And then from there, LinkedIn, Elm Street, Sean, hit me up. Awesome. Thank you so much. Before you leave, if you're interested, here are three important things that happened in the AI world in this past week, summarized into a few short minutes. The two first items are tied together, I think. One of them is whether you know that or not, back in the end of March, Italy banned the use of ChatGPT.
42:20And they banned ChatGPT on concerns on customer data privacy and GDPR and how they've collected data and how they're using data. And they went back and forth with OpenAI in order to get them to make changes to the way they work. I seriously doubt they got them all the way to the point they wanted. But the bottom line is this past week, they removed the ban and saying that OpenAI has agreed to take the right measures in order to protect customer data. My take on this is that there's a lot of politics going on in the background. This obviously first step by Italy projected into what might happen in all of Europe because GDPR is just an Italian rule.
43:02It's a European rule. And I think when I say politics is I believe what happened is that OpenAI agreed to take several measures. We're going to talk about one of them in a minute. and I think European Union, as well as Italy specifically, understood that this train has left the station and that there's a lot of other platforms out there so they can ban ChatGPT, but I don't think they can ban all of them. And then what will happen is that the Europeans will be at a big disadvantage compared to anybody else in the world that can use these tools. So they found a way to compromise and allow ChatGPT to be used with taking several different measures that they've agreed on.
43:41So that's my take on this. And if to go specifically to the second update in ChatGPT today, there's a new settings where you can go in and toggle between allowing it to use the information you're putting into ChatGPT to train their model or not, which means at least on future conversations you're going to have with ChatGPT, you can have it not use the data you're putting into it if you're interested in doing so. So that's probably connected to the other piece of news. The third piece of news also related to OpenAI is that in the past few days, they've rolled out plugins to a lot more people. And so more and more people now have access to ChatGPT with internet access and other plugins for ChatGPT.
44:24Those of you who don't know what plugins is, are different capabilities that are like add-ons to what ChatGPT can do on its own. Some of them are absolutely mind-blowing. And you can sign up for that for a wait list and then eventually we'll have that as well. if you want access to that yourself you can go to openai.com forward slash waitlist forward slash plugins and you can sign up and then like i said they started rolling this out to more and more people so you'll have access to even more capabilities than you had so far that's it for this week stay tuned learn a lot about ai use it as much as you can share with me on linkedin anything you find or anything you want me to cover on the show and have an amazing week until next time
From the publisher
What if you could tap into the limitless potential of AI and automation to revolutionize your business processes?
In this mind-blowing episode, we explore the incredible world of AI-powered automation and its transformative impact on businesses. Our guest, Shawn David, a seasoned expert in automation and AI, shares his wisdom and insights on how to embrace these cutting-edge technologies to skyrocket your business to the next level.
🌟 Topics we discussed:
🚀 Understanding your own processes and identifying repetitive tasks
🔥Identifying and automating low-hanging fruit tasks
🛠️ Leveraging the power of AI for next-level automation in replacing manual tasks
🤖 Using AI to assist in identifying more opportunities for automation tasks, optimize processes, and even create custom solutions
💡Utilizing no-code tools like Zapier for automation
🌐 Incorporating scripting as a low-code solution for Automation
💼 The future of AI in various industries and the potential for rapid change
🔮 How AI can revolutionize businesses, from creating new apps to transforming the way we work
Shawn David is an AI and automation expert, helping businesses unlock the potential of these powerful technologies in Automate to Win. With a passion for innovation and a knack for creative problem-solving, Shawn has made it his mission to empower businesses with the tools they need to thrive in the fast-paced digital world.
About Leveraging AI
- The Ultimate AI Course for Business People: https://multiplai.ai/ai-course/
- YouTube Full Episodes: https://www.youtube.com/@Multiplai_AI/
- Connect with Isar Meitis: https://www.linkedin.com/in/isarmeitis/
- Join our Live Sessions, AI Hangouts and newsletter: https://services.multiplai.ai/events
If you’ve enjoyed or benefited from some of the insights of this episode, leave us a five-star review on your favorite podcast platform, and let us know what you learned, found helpful, or liked most about this show!



