Transcription
I'm going to show you how to set up a no code team of AI personal agents that can do just about anything, like getting your calendar, email inbox, updating your CRM or project management tool, sending emails, and even creating meetings and events in your calendar with desired guests.
So, by texting my agent on Messenger or even having Siri send a voice message or text, I can have this agent summarize my schedule for today, tomorrow, next week, next month, maybe summarize the emails I've received from yesterday or even today, and really anything to do with personal administration.
The reason I built this is because I recently started working more with a no code agent building platform called n, and I wanted to try out something a little more complex with the software. Hence the reason I'm building this out now.
The task that this agent does for you really can save you a lot of time in the long run and, more importantly, helps you stay informed. I'm going to show you how to build this out step by step in this video. But real quick, I wanted to mention that if you're a business owner and you're tired of being stuck in the day-to-day grind and want to leverage our systems to scale your business and completely eliminate your repetitive tasks, then book a free call with me. I'll show you how to drive revenue, save time on your business, and ultimately scale for growth so that you can finally step away from the daily operations.
Also, be sure to join my Skool Community because I'll be providing so many other valuable resources like this automation that will help you grow and scale your business, which I provide on my YouTube.
So, this is what the automation looks like in N, the platform that I use to build out this entire agent. It's pretty robust, but I want to show you how it works and showcase some of its abilities. It could go down every one of these different paths: getting calendars, sending emails, getting emails, creating meetings, and updating my project management tool.
Let's say we want to get our emails from yesterday. So, I'm using an iPhone; I'm just going to voice activate it. I don't want to feel like texting out the agent, so I'm going to say something like, "Hey Siri, text my AI agent. Can you get my emails from yesterday?"
Send. It just sent this off, and in a second, I should be getting a message. So, there's my AI agent. Basically, it's just repeating the message that I sent off and then saying, "Here's your email." I got two emails from yesterday, and as you can see, they're listed out right here.
For one, I included who it's from, so there's a little "From" section and there's also a summary section. But it doesn't stop there. As you can see, I have this entire list. Let's call this my CRM. I just use ChatGPT to create this for me.
Let's say you have a giant list of clients or leads, and you want to input them into your database or need to put them in your database. Basically, I gave N access, or my AI agent access, to this sheet. So, let's say these are just some of my friends or something like that, and I want to send an email to one of them or maybe I can even create a meeting with one of them.
So, what I can do is I'm just going to say, "Can you create a meeting with Alicia Boyd on Thursday at 1 p.m.?" I'm going to send this off. It's going to repeat the message, and if I check my calendar, so if I go to Friday, I'll just refresh, scroll down, and go to Thursday. As you can see, there's the meeting set up: Alicia and Nick synced. It just created that meeting automatically for me.
It found this person in my CRM and has access to my database. If you have a vast knowledge base or anything like that, you can give it access to this AI agent. There are a lot of different use cases there.
But let's say we want to send an email to Brianna Miller. So, if I pull it up, I say, "Send an email to Brianna Miller just telling her that we're unsatisfied with the services that we've been getting recently and we are going to go to your competitor."
I'm going to send that off. As you can see, I used voice activation once again, but also, you could just text it out. It's going to repeat the message and then it's going to say, "Done. Was there anything else?"
What I did here is instead of actually sending it off, I made it create a draft because I just want to review it. So once it says done, that signifies that I should go review it. If I refresh and come into my email, as you can see, there's a draft right there. It says "Unsatisfactory Services Feedback."
So it created a subject line for me and compiled this entire email based off of what I said. It says we're unsatisfied with the services that we've been getting recently, and you know exactly what I said in the message.
So, it could do so much more than that. It can, you know, get your calendar. If I wanted to, you know, list out my calendar for next month or something like that, then it could do so. If I wanted to list out my meetings for tomorrow or if I have any meetings for next week, it can do that.
If I want to update my project management tool, so if I want to create a task, you know, a lot of times I'll create tasks in my Notion page. Basically, I'll just say something like, "You need to write down your five beliefs" or "do 20 push-ups," I don't know, something random.
But you can have it do that. If I wanted to do that, I'm just going to pull up my phone again and going to say, "Create a task for me to do 50 push-ups on Thursday this week." I'm going to send it off. It's going to repeat the message, and if I come into Notion—let me make sure I can find it.
I'll go to my project management section once this loads in. If I scroll down, you can see one of the tasks is "Complete 50 push-ups," and this is for tomorrow. So, if I were to open my calendar, I'll find this for tomorrow as well.
But let me go ahead and actually open up this page. As you can see, it listed off as tomorrow, and you could even put in more credentials. I could change the importance of this; I could change the time to complete it. But as I did right now, it's just the execution date.
So, there are so many different things that this can do, and this probably took me, I think this took me like an hour to actually set up. If you already have an understanding of make.com, then you should have an understanding of N. It kind of goes hand in hand and it's pretty easy to understand.
So imagine if you were to spend, you know, more time than that. If you were to spend a few hours or a week, or a couple of weeks building out some systems, then you could have something that's pretty insane that can just act as your administrative assistant.
So maybe you don't even have to employ someone, you know, paying them $40,000 a year, $50,000, even more, to do these administrative assistance tasks that this AI agent can take care of everything.
So anyway, that's enough rambling. Now, I want to get into how we actually built this out and how you can do so for yourself. This looks pretty complex, but I think that once I break it down and show you how I did it step by step, it will help you achieve the same results as I did. Whether you're using make.com, Zapier, or N, it doesn't matter; they're pretty much all the same across the board.
So we're starting off with a trigger, and a lot of these different platforms have the same modules and everything. So feel free to follow along on a different platform. But anyways, what we're using is Twilio.
So what Twilio is, is a platform that allows me to buy numbers and deploy them on platforms like this. Every time I'm saying, "Siri, can you text this person?" I'm texting my AI agent with the number that I bought from or I'm sending it to the number that I bought from Twilio.
So let's say the number that I bought is, you know, 856. I'm texting that number 856, specifically with the directions and everything. As you can see, we're starting off with this trigger.
So all I did was buy that number. You could buy one for, I don't know, 30 cents, a few dollars, something like that, and then you're just connecting the account from Twilio. So you're finding the API key and everything and you're just plugging that in—simple stuff right there.
What we're doing then—and basically this trigger—is just watching for messages. If we come into here, we go to add a node, and if we search Twilio, you'll see it here. Then, you'll just find triggers right here. On new SMS, you can just input that right there.
Want to delete this? But once we receive that message, what do we want to happen? Well, we want to do all these different things that we built out. So to get there, we're going to start with an OpenAI module.
I just want to repeat again that this isn't the best practice to go about everything. I'm relatively new to using N, so I know that there are different functionalities like AI agents and everything which I'm still exploring. But this is what I've learned and how I got the project across the finish line. Just know that there are practices for doing this whole thing, but this took me just an hour or so to build out. It works for me.
Now, let's zoom in a little bit and go to OpenAI. What we're doing is we're taking that text message that we just received and we want to analyze it. We want to see what it was saying and give it some specific outputs from there.
So I'll read off what it's saying: "Your task is to analyze the message and decide which actions to take. There are five possible actions, and your output should be based on the content of the message and apply common sense to determine the correct actions. You must ensure that your output only includes relevant actions based on the specific message provided. The five actions are getting your calendar, getting emails, creating tasks, sending emails, and creating meetings."
So as you can see, the output over here on the right side indicates the output. Here's the input: the content or the result is going to be to get emails and send email because in the message I asked basically my AI assistant to get emails for me from a specific day and to send an email to a specific person about something.
So it decided to go with these two outputs because that's basically what I was asking for. I was just breaking it down into two very simple binary things. So I included this action. Important: your output should only contain the actions that are directly relevant to the message.
For example, if the message suggests retrieving emails, the output should be "get emails" and nothing else unless other actions are clearly indicated by the message content. Then, I provide it with some further actions.
This is so important when it comes to prompting: providing your AI with some examples. So let's say you don't give any examples. I mean, your AI probably will still work properly, but it won't work as well or as often.
So you're just ensuring that the conversation is going to be 100% more effective. Just giving it some examples of what it should look like properly. So if the message says, "Please send me the emails from last week," your output should be "get emails."
If the message says, "We need to set up a meeting for tomorrow," your output should be "create meeting." If the message says, "Can you check my schedule and send a follow-up email?" your output should be "get calendar, send email."
Then the instructions are to analyze the message content. Then I'm giving it the message content here. So here's the JSON. All I did was plug in the body from all the way down here and drag it into there.
So then I said, "Use common sense and logic to decide the correct actions based on the message. Your output should include only the relevant actions and nothing else. If the message implies only one action, I'll put just that action."
So I give it the message that it's analyzing and everything. That's one of the most important steps of this entire thing. If you can understand that step and this step, then you kind of understand the entire thing.
I don't want to say you're halfway there, but you're halfway there to understanding how to do stuff like this. So let's say we get our actions. So what was our output in this example? "Get emails, send email."
So with that, it's going to go down this route: get emails, send an email.
So how did we achieve this and ensure it doesn't go down all these different paths? I forget what this path was right here. So let me make sure I just rename this so it's easy to understand. So this one was updating project management.
All these different tasks we have: getting email, sending email, getting calendar, updating project management or updating task or creating task, whatever. Then creating meeting where you're just sending out calendar invites and all of that and Google Meet invites.
So how do we make sure it doesn't go through all of these? Well, that's something that we use called filters. So what we did is we opened up—sorry, not opened up—we added a node and we just searched "if."
So if you search "if," this is where you're going to route different things and you're going to be able to filter different stuff. Let me delete that and open up this. So this is the filter right here.
If my message content from ChatGPT contains "get emails," then it's going to go on this route. It's not going to go down other things. It's going to go on this route specifically, and it's not going to allow anything else to come in here unless it contains specifically this wording.
As you remember in ChatGPT in this module right here, we prompted it to say different things like getting emails if we're asking to find our emails from a day or whatever. So that's essentially how the filter is broken down.
Next, what we're going to be doing is finding the day that we're talking about getting our emails. So let's say I asked my AI assistant, "Hey, can you get my emails from yesterday?" Well, we have to know the current date and we have to know what date they're talking about.
We may understand it, but we have to make sure AI and our assistants are understanding it. So how we did that is we used a tool called date and time. You can find this by simply searching it.
If you search dates and you find "get current date," I inputted that right there. So bringing this little node and inputting it into date and time, then connecting it to ChatGPT.
Now, what this is doing particularly is it's extracting the content from the body and it's determining the day to get. So here's what I said particularly: "You are to determine the date to get. If the text message asks to get emails for today, then your output should be today's date. For example, today's date is the 10th of September, then your output should be 2024 September and the day of the month, which is 10. Time doesn't really matter; we could always just keep that as zeros.
For example, if the text asked to get emails from yesterday and the date is the 10th of September, then your output should be the 9th of September 2024. So for example, if the message asked to get emails from last week and today's date is the 17th of September, then your output should be the 7th of September.
Then I just give it the context; I'm giving it the message content and I'm also giving it the current date. Your only output should look like this, and it should never say anything more.
So let's go ahead and test this out. I'm not sure if it's going to give a proper example. Okay, it will. So what did I say here in the message? I think I said to get emails from yesterday.
So I said, "Hey, can you get my emails from yesterday? Also, send an email to Nancy Curry just saying that I want to meet up for a coffee tomorrow at 4." So let's focus on this first part. It says, "Hey, can you get my emails from yesterday?"
So, today is the 11th; it's getting the 10th. So as you can see, there's the 10th right there, just as I asked it for. Up here, we're connecting it to a Gmail module.
So as you do search Gmail, get message, get many messages, and you'll get exactly what I am looking at in here. One second, let me delete that.
So you'll get this, connecting you to my Gmail account. I'm getting many messages from this operation. I'm only returning 20; I'm probably never going to return more than that because I don't get very many emails every day.
Don't have to focus on this; just keep that simplified. But what we are focusing on is the filters. So label names or IDs—so how you get that exactly. So if I go to “Add filter,” and I'm basically—hold on one second—so if I were to delete this, label names or IDs would come up, same with received after.
Once I get the label names or IDs on here, I'm just searching for inbox and the unread because those are the only two that I'm looking for. I don't want anything. I don't want it to give me anything in my drafts. I don't want it to give me emails that I've sent. I don't want it to give me spam; that won't be my inbox. I only wanted to get my inbox and unread emails because if I've already read them, then I probably know what's going on already, and I don't need to get notifications about them or summarizations.
Then, I'm getting received after, and this is just where you're putting the previous AI modules' content, so I just drag this in here. So I'll even show you; just drag it in and put it right there, and I'll test it out. In a second, it should output those emails.
So here's the two emails that I got from yesterday, from the 10th of September. In a second, it's also going to text me that. So let me make sure I don't screw this up. So why it's going to text me that is because I then sent a Twilio or created a Twilio node.
So this node right here, as I showed you earlier, basically you're just adding a node, search Twilio, and this specific action is just sending an SMS message. You can also use WhatsApp, but I just like using Messenger because it's so accessible, and I can use Siri to do it for me.
I can literally be in the car and just call out my phone and be like, "Hey Siri, I just need you to text my AI agent whatever." But basically, I'm just including the context.
So I'm saying, "Here's your emails," and from the previous node, I'm including the JSON. So that would be this snippet right here. I'm dragging that and putting it right there. I'm sorry; I'm dragging the 'from,' so that would be right here. So dragging this and putting it right there in the summary.
That's where I'm grabbing the snippet, so that's obviously the summary. And that is all I'm including in the message. So the 'from' is obviously the number that I have hooked up to Twilio, and the 'to' is just the number that you want to send it to. So I would recommend your personal number.
Cool, so now let's move on to the next step. So let's say we want to send an email. If we want to send an email, let's say I'm sending it to Nancy Curry, because if we come into our CRM, Nancy Curry should be one of the numbers in here.
So yep, here's the email and everything. So here's the company, fake company, Mendoza Cook and Williams email, first name, last name. So we’re going to send an email. Obviously, we have to find out their emails and everything.
So let's say you have your database and you have your database containing a bunch of clients, leads, friends, contacts, whatever you name it. It's this mass database, and you want to give it context to your agent. You just want to train it on all that information.
So anytime you're trying to look for a specific person, say you're trying to look for Mark Williams, it's going to search in my CRM for a Mark Williams, get their email, put that into Gmail, and send them an email.
So how do we go about doing all of this? As you can see, I requested to send an email to Nancy Curry. So now we have this filter; it is going to go down this route. So if my message content from the previous module contains "send email," then it's going to run down this route.
So, very similar to what we just discussed previously, it's just going to do that for the sending of emails. The first thing we want to do is get the last name. Sometimes there are going to be similar first names, even last names sometimes, but it's a lot less common. Hopefully, you're going to differentiate between those contacts if they do have the same last name somehow.
But essentially, what I'm going to do to get the last names from here is look at the message and just give me the last name. Only make that your output. So, for example, Morrison. I gave it that output because the message contained Mark Morrison.
Here's the message and the body of the message, basically. If I were to test this, I believe I said Nan Curry, so the output should be Curry. So if I scroll over to the content, if I go to the expression and now test this out, as you can see, my answer is Curry.
So it didn't spell the name properly, but that's okay. But anyways, once we get that last name, we should be able to find it in the next module. So what we're doing is a Google Sheets module.
Let me zoom in a little bit. With Google Sheets, basically, I'm just connecting it one second. So connecting it to the action. Google Sheets, what I'm doing is get rows and sheets, and then you'll get this little node right here.
So let me go ahead and test this right now and see if it finds the last name. So it's not going to output it properly because it didn't get the—didn't spell the name properly. So sometimes you might have to specify, but in this specific case, we're just going to type in Curry and then test the step, and then we should get our output.
So, because we have a Nancy Curry in our CRM, so right here, it then outputted it, and it’s giving me that row number from Google Sheets: the first name, last name, email, and the company.
So this is what we need to then give to our next step, which I will show you in a second. But let me go ahead and replace this value before I forget to do it later. So let's go to content, replace that in there.
And next, you want to create the email body. So let me go ahead and expand this. So what I'm saying here is you are to look at the message and determine what is to be in the email body. So for example, if the message says "send an email to Mark Morrison asking for their cell phone number," then your output should be "could you send over your cell phone number?"
So you should use common sense to determine what should be in the body and not add or remove anything from what was mentioned in the message. You can just say it word for word.
Here's another example: if the message says "send an email to Matthew Curry," I guess I could say Nancy Curry, and your output should be "we just sent payment and your check should be mailed out to you within the next few weeks. Be on the lookout for it."
Here's the messages to analyze, and then I'm just giving it the Twilio message body. So let's test this out, and basically it's just outputting "I want to meet up for coffee tomorrow at 4."
So it sent that email because if I go to the message body, I said, "send an email to Nancy Curry just saying that I want to meet up for coffee tomorrow at 4."
So I could have proposed it in a nicer manner or something like that, but this is just how I said it particularly. So now what do we want to do next? Well, now we want to basically just template this and make it look nice.
So basically, I'm just including formatting: including a hey, inserting the first name, inserting an outro, just saying like "best Nick." And yeah, so this part I won't go too deep into, but essentially I'm just saying an introduction should always include "hey," insert the first name.
So I'm grabbing the—if I go to Google Sheets, go to the schema, grab this first name, then put it right there. If I were to test this out, I will see what the email will look like.
So message content is going to be something like "Kenan, I want to meet up for coffee tomorrow at 4. Best, Nick." So I think it looks pretty solid. Now, there's just one more thing: we need a subject line.
So let me go ahead and expand this as well. So you were to determine a subject line from the message. It can be simple and just contain a summary of what the body is about. Keep it brief, simple, and easy to understand.
Here's the body you are basing the subject off, so I'm giving it the body items and everything, some of the message content. Then I say if the body contains something like "we are canceling our membership due to unsatisfactory service," then your output should be "canceling our service."
I mentioned that; that’s just an example. So now let's test this out and see what happens. I have to change this; make sure it's getting the schema properly. The content is going to be "meeting for coffee tomorrow at 4 p.m." That's going to be the subject line.
Now, let's go ahead and create a draft of this. So what I did here, went to Gmail, searched up Gmail, searched draft, so create a draft. And here's what I get here; accidentally executing it right now.
But anyway, this is what you'll get. Hook up your account as before; make sure the resource is the draft and that you're creating it. So the subject line—so as you can see, this is the previous module where it's creating the subject, drag this in here.
And then here's the body, so this is going to be the email body right here. Input that right there. And then who it's actually going to be sent to, so if we go to Google Sheets, grab the schema from here and the email, just drag it in there.
So we'll test this out once again and we're going to check our email draft, and it should be in there somewhere. So if I refresh this, okay, so "meeting for coffee tomorrow at 4 p.m."
So it's everything that I asked for properly; here's the email. All we have to do now is just review this and send it off. If you want to make changes, you can make changes, but what you can also do is just make it send that actual email.
Instead of just creating a draft, you can have it actually send it, or maybe you can even create a step in N where you're just saying it's asking you to review, and once you confirm that it looks good, it can send it for you as well.
So there are a bunch of different steps that you can take. And like I mentioned, I spent just like an hour or two hours at most building this out.
So take your time with this. You can do some really cool stuff if you're actually trying to build something valuable. Now, what we're doing next is, this is basically just a message confirming that we've done this.
So basically, in the message, I'm just saying "done, was there anything else?" So all I did here was just creating a node for Twilio, similar to before. I'm just sending an SMS message to my number and everything and sending that off once everything is good.
So that takes care of the first two rows which deal with emails—getting the emails, searching them, and also sending emails.
Now let's go to the next route, which is getting your calendar. So let's say I have a very busy week next week, but I want to know specifically what the events are that I have coming up.
So we can do that. I don't think I messaged to get calendar in here, but if I were to text my assistant, so I'm going to open up my AI agent. Actually, let's do it like this: "Hey Siri, can you text my AI agent saying can you get my calendar for the next two weeks?"
I'm going to send that off and in a second it should come up. So as you can see, it ran through all of this. Now it's going to get my calendar.
So as you can see, it's going through check marks of all these different stages, but it's not passing because it should only contain G Calendar.
So let's make sure that it says G Calendar. So let's go here, test the step, make sure that it's in the expression so that it has context to the message. And the message is get calendar.
So now it should run through this little stage right here and go all the way down here. So now we want to start building out the getting of the calendar.
So similar as before, we have our filters. So we're giving it a message content from the previous ChatGPT module, making sure that it contains "get calendar." And next we’re just giving it a date and time.
You could probably figure this one out for yourself, why you need to get the date and time. This is because you're trying to analyze when to get the schedule for, when not to get the schedule for. Because if I say get my calendar for the next two weeks, then I only want it to be for the next two weeks—not yesterday, not today.
So once we get the date and time, we're going to be giving it to ChatGPT, and what we're going to be saying—let me open this up a little bit—is you're to determine the date to get.
If the text message asked to get the calendar for today, then your output should be today's date. For example, if today's date is the 10th of September, then your output should be 2024, 10th of September.
And I'm giving it some more examples and everything, so feel free to look at what I'm saying and copy and paste it straight into your prompt. Make sure you change it back to expression so that it is formatting this context properly.
Close that out, and let me go ahead and just run this test real quick. So I’ll test this step. So I may have to run this whole module again. Okay, now.
So I did say get my content for the next two weeks, so the content is the 25th. If I were to check my calendar, the 25th is Wednesday in two weeks. So that's good.
Now we want to do is get before, so making sure to exclude any other dates. So what I'm saying here is you are to look at the output of this message and determine the end date. So I'm giving it the Twilio body message, the original message that I'm sending to my AI agent.
Then I'm saying you will have to analyze the message and determine an end date. The message could say many different things including, but not limited to, today. Maybe it says for Wednesday of next week; maybe it could say Monday of the 20th of September 2024.
I’m giving it a bunch of different examples and all that, similar to before. But let’s go ahead and test this out. The biggest thing about this entire automation is it just comes down to prompting.
Yeah, so if you follow my steps, you should be good. So as you can see, we're getting the date right here, so the 25th. Now let's move on to the next stage.
So Google Calendar—basically, I’m including the after and the before. So dragging the after, so I listed it out so you can easily understand it.
So the content of the after, dragging it into here; before, dragging it into here. We run this and it should list out my calendar. So I don't have anything listed out for the next two weeks, so we're good. But if I were to create a calendar—and I guess let's do it for the 25th—then it should list up something.
So it'll say next date; we'll save that and we'll try running this again. Go to Google Calendar, test output.
Okay, cool. So now let's move on to the next stage and actually send out any of those calendars or any of those events to my number.
So in this next step, what I'm doing is creating another Twilio, just sending a message where it's going to list out the events and everything.
So in the message, I'm saying, "Here's your event schedule," then I include the summary from the previous module—the JSON summary and the JSON start date, which is just when the event is going to start and everything.
So that's how you build out the getting of your calendars—pretty cool step right there. And now let’s do the updating of my tasks in Notion.
So now let me open up my phone. I'm going to say something different. What can I say? So can you create a task for me that will tell me to buy a whiteboard?
So before I send it off, I want to make sure to test our workflow. So now I'm going to send it, and in a second it should trigger. So here we go.
Now it's going down this entire module. So why did this happen? Well, because it determined that I am creating a task.
So now if I go to my Notion—so, I don't have it open currently. So I go to Notion, I'll go to my to-do list, which is where I basically handle my project management.
So if I search for buy a whiteboard, here it is. So I didn't specify a particular date, so it just said tomorrow. So I just have to make sure to include a date, so that's a slip-up on my part.
But yeah, if you don't include the date, AI’s just going to randomly generate one for you and just put it in your project management somewhere.
But it shows to create a task because in this, my filter is creating a task. If the message contains "create task," it will thus go down towards this route.
So next, we're determining or getting the current date and time once again. So as you can see, there's a pattern between everything here. Everyone is including the current date and time, so I probably could have included it right here.
So I just include one and then it goes on forever, and I don't have to do it a bunch of different times. But yeah, so I did this kind of on the fly—built it really quickly as you can tell; everything's sloppy and not very presentable.
But it works for me, as I said. So next, if we want to update our project management, we have to find out the date and everything.
So if I said get my task for or buy a whiteboard tomorrow, then it's going to mention that for tomorrow or put the execution date for tomorrow, should say.
So in this prompt, what I'm saying right here is your job is to look at the message and determine what the name of the task should be. It should be concise and never more than 10 words; even 10 words is a lot.
Then I'm giving it the content or the context of the message, so including the Twilio message. So the text that I sent to the agent, and I’m giving an example.
So if the message says something like "create a task to send a voting ballot on Thursday," then your output should be "send voting ballot," and only that.
I wanted to make it very simple. So as you can see, I mentioned I want to buy a whiteboard. Then it just simply says buy a whiteboard.
Cool, and we'll go down to the next route. And basically, I'm saying here your job is to look at the message and determine what the execution date of the task should be.
So here's the message, giving it the context of the Twilio message, and here's the current date. Then I'm giving it some more examples.
So now it's providing the date. So I didn't—let me make sure I messaged it tomorrow. Yeah, so I didn't message tomorrow, but if I were to say tomorrow, then it's going to come up with tomorrow's date.
So today is the 11th; it would have listed out the 12th. Let me go ahead and do that real quick just to prove that this is right.
Create a task for me to buy a whiteboard tomorrow. So I will send that off. It should go down towards this route and we'll find another one.
So let me delete this one; delete a couple of things. So buy a whiteboard. So there it is. Now the execution date is for tomorrow, so working properly.
Cool, so now that we have all those props listed out as we want, then what we're doing next is hooking up our Notion table or our database.
All I did was search Notion, and you just had to integrate or create an integration in Notion. So you do that by going here and just managing your connections, connecting this to N and everything, and yeah, just making sure that it's properly integrated.
So what I did here is, I believe it was create a database page. Yeah. So as you can see, create database page. If you open that up, you'll get this.
So for the title, I'm just including the message content so that would be, I believe this, so including that context and then just including the date so that will be the execution dates and then this will create the output for you.
Let me show you the output. So it’s giving me this right here, so as you can see here’s all the properties: the status, time to complete, execution date, name— all of that.
Then I'm having a Twilio message. Basically, I'm just saying at the end, in some way said before, I'm saying specifically "done anything else?" So just allowing you to ask more questions.
Next up is the creating of meetings. So let's say I want to schedule a time or, you know, create a Google Meet with, let’s find somebody in our calendar or in our database.
So we'll do Todd Nichols. So let's say we want to schedule a Google Meet with Todd Nichols for next Tuesday. So today is September 11, and next Tuesday would be the 17th of September.
So I'm going to open up my workflow; I'm going to test it out. So what was the name again? Todd Nichols. And I believe I mentioned Tuesday.
So I'm going to say, "Can you schedule an event with Todd Nichols for next Tuesday? The event's going to be about making money." I'm going to send.
So, opening this up, it's going down this route because our filter was to create meetings. Similar to before, it's all super similar—basically the same exact thing.
So going down here, getting the date and time once again. So we'll do here.
So similarly to the other modules, we have a prompt that’s basically going to the date and time. So what did I say here? I said I wanted to get the event for next Tuesday.
So it's finding the date of today: today was the 11th of September, and now it's finding when Tuesday would be. So that's the 17th, as I mentioned before it would be.
So I’m just mentioning the some things to include and giving it a bunch of examples. So make sure you prompt your AI properly. So I'll even scroll down so you have a chance to copy this and write it out.
Okay, you can simplify the output; we're not really too worried about that. But the next step is to find the end time.
So we want to make a meeting for probably an hour. I'm always going to make the meeting for an hour. So for the end time, I'm just looking at the last message or the last output.
So I'm looking at here and then just adding an hour after, so I didn't mention in the message what time I want the meeting, which is probably pretty important. And it is important.
So I'm going to restart this and I'm going to say, "Can you schedule an event with Todd Nichols for next Tuesday at 11 a.m.? The event is going to be about making money."
I'll send, and now it's going to run through this again and it should be right this time, I hope.
So if it's going to extract the date and time, it should be 11 a.m. So let me make sure I have it set to expression; test that out, and we should find that it's 11 a.m.
So yep, so the 17th at 11 right there, so the end time should be 12 in here because it's just going to add on an hour. Our events are always going to be one hour exactly.
So if I test this, okay, as you can see noon right there, and now for this part, we're just going to get the last name because to create the event, we have to create the meeting, we have to create an attendee, and invite them to it.
So to do that, we got to find their email from our database. So we're just plugging our database into our agent, giving it that context. It's going to find everything from that database, find Todd Nichols, get the email, and create the event with them.
So for here, I already created this module before, so I'm not going to show this again, but I'm just getting the last name.
So once I get that last name from ChatGPT, I'm going to create a Google Sheets, same as before, for where I'm just going to search the entire CRM.
So I'm getting the rows, finding my sheets, connecting it, giving it the value of the name and everything. So let me go ahead and test this out real quick.
So Todd Nichols is right here; this is the last name I put. I put the last name right in here, so I just dragged it, put it in here, and it came up with Todd Nichols.
So row number 58; here’s the email. If I go in here, I should find row number 58. Yep, so perfect—just as we wanted.
And next, it's going to create a Google Calendar. So to do this, add a node, go to Google Calendar, and what I'm doing is just creating an event.
So there's a bunch of different options like getting your availability and stuff like that, as you just saw. So lots of different use cases for this, but basically we're just including the start time, the end time.
So the start time—I don't think I have it listed out. No, I don't. So it's hard to understand which one’s which. But great meeting? No, so I'm not sure which one it is.
Okay, I think it's date time to get so 11 a.m.—okay, yeah. So the meeting is starting at 11 a.m.
So AI or ChatGPT gave me that output. I'm dragging this output into the start—that's when the meeting is going to start. When it's going to end, we created a prompt or an AI model for that, so we grab this content and put it into the end.
And now we want to include additional fields. So for the additional fields, you’ll just go to add a field, click attendees, and also summary.
So the summary is going to be the title. Attendees is where I'm going to include the email. So if you go to Google Sheets, as you can see in the schema—let me close out some of these—go to the schema, go to Google Sheets. You'll find the email and input that right there.
So that'll make sure it’s going to send the email to them that the event was created and everything, it's going to send them reminders and all that good stuff.
So for the summary, basically all I did is just include their first name. So I don't want it to be anything too crazy.
To get their first name— as you can see from Google Sheets— all I did was just drag it in here and then I said, "and Nick synced."
So now let me test this out and I'll show you what it should look like. So if I run this and if I go to my calendar, I believe this is for next Tuesday.
So here's Wednesday. I'll refresh this, make sure I'm going to the 17th. So it should be with Todd Nichols on the 17th.
So there it is at 11 a.m. to noon. So just 1 hour, and it created it with Todd. So you can include more information. You can include descriptions, send more reminders to them through here.
There are so many different routes and use cases that you can go with this, and this is really just basic.
So this is just an administrative assistant that can do simple stuff. Imagine what else you can do.
Next, what we're doing is just sending an email and we're just saying, "Done, was there anything else?" So let me back out and show you this entire thing.
So this is what it looks like, so you can clean it up however you would like, and you know, feel free to use different platforms like make.com because it's a little more user-friendly.
This one isn't as definitely not as pretty and aesthetically pleasing, but yeah. So I know that was a lot, and I thank you for sticking to the end with me.
But yeah, so this was the AI agent personal assistant that I built out that handles an array of different tasks, and I'm going to clean this up a little bit more and actually employ it for myself and my company.
So you know, I can get my emails from yesterday, I can summarize all my emails, I can get my calendar, my availability, I can send out emails, and all that good stuff.
So there are so many real use cases that you can use for this automation that will help you save a lot of time because these little tasks do stack up, and it will end up saving you a lot of time and provide you with a lot of value in the long run.
But yeah, with that being said, I really thank you for watching, and I encourage you to book a call with me if you're looking to streamline some of your tasks or implement something like this for your own business.
Once again, thank you for watching, so feel free to book a call with me and my company using the link in the description below where we'll go over your biggest pain points, identify your bottlenecks, and prioritize things that we should be automating and actually doing it for you.
We could be building out things like this specifically from this video for your company to help you save time, provide more value, make more money, and leverage these systems to get a head start within your industry.
So you can find a link down below for that in the description. Also, I encourage you to join my free Skool Community because I provide you with resources like this automation that I don't give on my YouTube channel.
So with that being said, thanks for watching, like and subscribe, and I'll see you in the next video.