Leveraging AI
Dive into the world of artificial intelligence with 'Leveraging AI,' a podcast tailored for forward-thinking business professionals. Each episode brings insightful discussions on how AI can ethically transform business practices, offering practical solutions to day-to-day business challenges.
Join our host Isar Meitis (4 time CEO), and expert guests as they turn AI's complexities into actionable insights, and explore its ethical implications in the business world. Whether you are an AI novice or a seasoned professional, 'Leveraging AI' equips you with the knowledge and tools to harness AI's power responsibly and effectively. Tune in weekly for inspiring conversations and real-world applications. Subscribe now and unlock the potential of AI in your business.
Leveraging AI
103 | AI Productivity Boost: Automate Workflows with AI in Zapier Central with Valeriya Pilkevich
🔥 Special Offer Alert! 🔥
Transform Your Business with AI
Register Here (https://multiplai.ai/ai-course/)
Ready to revolutionize your workflow and boost your productivity? Join us for an exclusive live webinar with AI and automation expert, Valeriya Pilkevich. Valeriya will walk us through practical, step-by-step processes to build bots and automate workflows using Zapier Central.
In this session, you'll learn how to:
→ Automate your email and calendar management
→ Streamline your content creation process
→ Leverage AI tools like Relevance AI and Mind Studio for marketing and analytics
Valeriya Pilkevich, a seasoned AI enthusiast with a knack for simplifying complex workflows, will share her expertise and insights. With a wealth of experience in automating business processes, she will provide actionable tutorials that you can implement immediately.
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/
- Free AI Consultation: https://multiplai.ai/book-a-call/
- 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!
Hello and welcome to another live episode of the leveraging AI podcast, the podcast shows practical ethical ways to leverage AI to improve efficiency, grow your business and advance your career. This is Isar Mehtis, your host, and we have an absolutely amazing show for you today. But before we dive into that, one more piece of really exciting update from us. So those of you listening to the show for a while, I know I've been teaching AI courses since April of last year. We're teaching two courses every single month, but most of these courses are private, meaning companies and organizations hire me to train just their people, meaning. It's not accessible to the public. We open a public course about once a quarter. So we had one in October and then one in January and then one in April. And now we're opening the July one. It's going to start on July 8th and it's four sessions of two hours with me as the instructor with amazing people, always that, that participate and engage and help each other. So it's a very powerful course. It's called AI business transformation course. For a reason, it literally helped hundreds of companies by now, really understand how to implement AI successfully in a business context. So we talk about strategy. We talk about systems and processes. We talk about people. We talk about the steps that you need to do in order to implement this. And it includes a lot of hands on training and implementation across multiple aspects of the business. If that's something that interests you. Joyce is going to drop a link in the chat. If you're listening to this after the fact, then there's going to be a link in the show notes for you to find as part of the celebration of the 100th episode, we're doing a promotion that we've never done before, which is 200 off with the promo code leveraging AI 100 with leveraging AI as Uppercase, all of it. So leveraging ai, uppercase 100 to get$200 off the course, we're probably gonna run this for about a week, and that's it. So if that's something that interesting to you and you wanna do this a sign up because there's limited seats, and B, use the Provo code because that's going to expire. Now. Two, the really exciting episode that we have today. So you know, Amma. One of the big promises that AI gives us is business efficiency. And part of it comes through using AI within a automation of some sort. There's a few questions about that. Number one is how do we actually do this consistently? So those of you who are trying to build different automations with AI knows it not always works perfectly only sometimes. The other thing is that you, Probably no, if you try to use GPTs is and when I say GPTs is the custom GPTs within chat GPT, which is how a lot of people are now creating these small automations, they have a lot of limitations that did not allow you to do. everything that you want. Now, there's ways to go around that and connect GPTs or just regular prompts through tools like Zapier or make, but Zapier actually did something really interesting not too long ago. So two to three months ago, they've announced Zapier central, which is a tool within Zapier that allows you to build really amazing AI automations With the biggest benefit of them is that the fact that they're connected to the entire Zapier universe. So 6, 000 applications that are already connected are connected to Zapier Central. Now, our guest today is an expert. And then no code automation, ninja across every kind of business automation that you can imagine. So Valeria Pilkovic is passionate, really passionate. You will see that in a minute about marketing and about productivity and her background is marketing and marketing analytics. So both the marketing side, but as well as the data and automation and processes side of this, which makes her. The perfect expert on that. She's already created multiple Zapier central bots to help her run multiple aspects of her business. And she's doing the same thing for many of her clients. And today she's going to share with us specific use cases, like how to run more efficient email management, calendar management, content creation, stuff that every business needs built with Zapier central, and she's going to share. All the details and the steps exactly how you can do the same. I'm personally really excited to learn that because I started. Tinkering with it, but I haven't really built anything with Zapier Central yet, so I'm really excited to do this. Valeria with a really long intro. I'm really excited to have you at the show. Welcome to Leveraging AI.
Valeriya:Thank you. Thank you Isar. Yeah. I'm really excited to be here and share with you. My experience so far, please feel free to ask questions. And, of course I'm happy to share every secrets and, maybe first before I start, let me know in the chat, maybe just type a plus if you ever experimented with custom GPTs. custom GPTs are like the custom versions of the chatbot, of the OpenAI chatbot, obviously that you can fill with your instructions and knowledge base and so on. Maybe the first question you might have, okay, we already have our custom GPTs, which work fine for us, but what's in it in Zapier central bots, it's a bot as well, right? and the answer is there is few, there are a few differences or actually fewer advantages of Zapier bots. And first one, Isara already mentioned is that Zapier integrates already automatically with more than 6, 000 apps. Which means that if you played around with custom GPTs, you probably have seen this actions part below. And you know that unless you have coding knowledge, it's quite hard to actually set it up, this action. And even if you do, there are some instructions, for example, for Zapier, how you set them up, the schema to be able to access your calendar or email. But for me, I did it. And It's really lagging or sometimes it just doesn't retrieve or sometimes it would retrieve the emails from, I don't know, maybe last month or calendar events from completely not relevant. And sometimes it just, it gets broken and you have to set it up again. So this kind of integration part with. other external apps. If you are no code, if you do not really know how to code, it was hard. And this is already the first one, which Zapier Central solves, because I will show you, you do not even have to set up the exact automation. You do not have to use the code whatsoever. It actually, you just have to provide the access to your account, any website, and it will, they, I will guess basically what. What you want or what you need in this exact scenario. So that's the first part. The second part is live data. It's, and it's a huge one. It's, it sounds okay, That's not such a big deal, but it is if you think about your custom GPTs, you put the knowledge base, either a PDF or Excel. And let's say honest, let's be honest. Sometimes it even has difficulties to retrieve this data. Sometimes it's also only retrieves like a few parts of it, but not actually goes through the data and retrieves what you really need. And. with Zephyr Central, you can just connect it to your live data sources. You can connect to Airtable, you can connect to Notion database, you can connect to your Google Sheets, and as soon as you actually update the data there, it will automatically be updated. It will automatically be synced. So these are the two probably Huge differences. And maybe the third one is that for custom GPTs, you would have to go back to the GPT in case you need something, right? You would have to trigger it with your message. For Zapier Central, once you set it up, set up the behaviors, I will show you, it just works automatically. Like for me, I set up a few chatbots, which I will share with you today for email management, calendar management. And I think. like relevant for everyone. Literally, it doesn't matter if you have your own business, or if you're a professional and work for a company, or if you, I don't know, like for anyone, really very important or very handy. And it just works on autopilot. You just set it up when you, every time when you have a new email, or maybe an hour before you have a new meeting, it just runs it automatically. You don't even have to look at it. Personally, I never even go to the Zapier bots. I just set them up once And they work.
Isar:Amazing. I want to pause you for, first of all, a very quick summary and then ask you a question that is very relevant. the quick summary is three benefits that you mentioned. One is the fact that it's connected to the Zapier universe, which is huge. The second is that the data is live, right? So it's not, I've uploaded the document and that's it. and the third was you don't have to trigger it. It kind of triggers like any other Zapier. So basically what you're saying, it's like the best of both worlds. it's a Zapier automation, but that runs smart because it has AI in the background. So this is, Amazing. The question that I have, what are the biggest limitations that I've seen, or two limitations that I've seen on GPTs that have a lot of them. So I'm, I have nothing against GPTs. It's not this, right? There's GPTs are awesome. They have their use cases, but two limitations that they have. One is they're limited to 8, 000 words, which if you're trying to write a really long instruction, you get stuck and then you can't continue. And then you have to figure out what you want to give up on, or maybe say, String different to automate, like there's, you, you basically get stuck. And the second thing is you're limited with the amount of data you can upload as far as reference information, right? Either by the amount of files or by the size of files. And then you get stuck. Do these limitation exist on the Zapier side as well, like on Zapier central.
Valeriya:that's a great question. I haven't, honestly, like from the instruction part of it, maybe I did not create such a long instructions, but if you think about it already, if you have, if you can put separately your, your knowledge base. And if you could put separately already, like the integrations, the different behaviors, because you always have to think what's the, maybe what's the point of using those different tools. And as you mentioned, custom GPTs, it's a great example for many different use cases. And for example, If you are already used to the interface of open AI, and if you need, if you don't need this integrations, and if you don't need it to work on the backend, if you like need to, I don't know, to analyze the specific data, which you upload, I don't know, maybe customer sentiment or something else, then you can then simply stay with the custom GPTs. But if you need those integrations, or if you have a live data source, which constantly updates, this is when I would go to Zapier Central. But so far, I Like I didn't discover that there was like a specific amount and at least that it is not stated on the website that there is a specific amount of characters that you can put in the instruction. And, obviously you can chat in this chatbot, window as long as you want. So it, it does not it does not remember the data, what you chatted about, I don't know, weeks before, but you probably don't need it as well. Again, because the use cases I use it with are. Rather completely automated. So I don't really go to the chat bot to chat with it. It's rather, it just works in the background.
Isar:Okay, awesome. Quick question from the audience. Danielle is asking, does it integrate with SharePoint, Adobe Cloud or OneDrive?
Valeriya:Yeah, it must, we can check it now together. So there are 6, 000 apps and these are huge. That's what I'm saying. Consider
Isar:everything in Zapier a potential connection. So I'll be really surprised. I was, I would have been really surprised if there would have been no. But let's dive in. let's do the magic.
Valeriya:Yeah, let me share my screen and I will show you a few use cases and then I think it will be clearer
Isar:by the way, for those of you going to listen to this as the podcast, one is we're going to share and explain exactly everything that's on the screen. So you will understand what we're talking about. but B, if you're not driving right now or On the walk with your dog and you can watch, you can switch to our YouTube channel. So it's a Multiplai AI YouTube channel and Multiplai spelled with AI in the end, instead of a Y in the end. And we're going to drop the link for that in the chat so you can find it as well. and the link is also on the show notes. If you are on whatever podcast player, you can find it that way. But if you, are on the podcast, that's perfectly fine. We're going to explain everything that's on the screen.
Valeriya:Thank you. so this is just a website page, Zapier Central. It's a separate product of Zapier, but you have to have a Zapier account. So if you already have a Zapier account, you don't have to have a, paid plan. You can just start with a free plan. Just register at Zapier and then, you go to this page, Zapier Central. You can start, by just clicking this button, Try Central Today. and you have around 400 actions per month, which is if you set it up smartly, it will be even enough for you with this free plan. And I will show you, I will share with you some secrets. And as you can see here, it's already, it's talking about the benefits like connecting the live data and 6, 000 integrations and so on. And the great about this site, which I really recommend you to check out is that it already gives you, some. Let's say some inspiration for the bots that you can create because obviously every one of us have different use cases and, you can just check it out and already there will be like a pre filled integrations or pre filled instructions. For example, yeah, pre event notifications, hot leap notifications, RSS bot and so on. Just check it out and, yeah, and start experimenting. So for me, let me know if you are able to see my screen now, actually with the chat. Okay. So I will show you, I will start with the two, which are like really run on autopilot and are really helpful to me. One is the email assistant bot I created, and the other one is calendar bot. So in Zapier, you probably already recognize the interface. It's a chat. And, the different thing from, for example, custom GPTs here is that again, data sources. So you can connect your live data sources and let's see what you can connect. So for live data sources, it's still limited what you can connect, but they always work on it. I think there are already more integrations than two months, before. So you can collect data from Google Sheets, even from Excel, from, so Google Drive, from Airtable, from Notion, which means that if you have the database in Notion, You just automatically connect to choose your Notion account. And once you update something in Notion, it will be synced here as well. But this is, do not confuse this with the integrations with the 6, 000 app where it actually integrates for, to perform different actions. Then you have instant actions. Probably, basically, if you stay in this chat and start chatting to it, for example, I already added some actions, like I already added my Gmail account and I allowed it to create draft and label emails and find different emails. So basically, I can already start chatting with it and say, okay, show me the emails that, that are incoming today and, I can start chatting with it. But. It's not the best part. The best part is the behaviors. So behaviors is exactly what you set up and what differentiates it from custom GPTs, which work then automatically. So for behaviors, there's usually a certain trigger. So for example, when this behavior starts, it's if you use Zapier or Make, you know what the trigger is. And, for example, currently I set up when I, every time when I received a new email and I'm using Gmail, but again, you can use Outlook, you can use, all the different emails account that is possible, Microsoft Outlook and so on. Or what I was telling to you about, if you don't want to use, or if you don't want to be run out of your actions very quickly, you can even set up something called scheduled by Zapier and you can set it up for example, to only check your emails once a day,
Isar:once a
Valeriya:day or twice a day, and it's maybe every time, because currently trans the action every time I get the email. And honestly, I'm subscribed to so many newsletters. I get the email like every five minutes, which it just sends to archive and I'm using my actions. So maybe it's not the smartest way. Maybe the smartest would be set to say, okay, just check it like three times a day or five times a day and just label them and do everything that you want. So again,
Isar:again, just to. clarify the difference. I assume it doesn't matter how many times you run it. It will still run on the X number of emails you have, but because you're going to have less emails on your inbox, because they're automatically get sent to archive. That's why you're going to save the money, right? It's not the fact that it runs less times. It's the fact that there's going to be less emails to work with.
Valeriya:No, it's actually, it will be considered one action. you just set it up. So even if it brings in like
Isar:35 emails at one time, it's still one action. Oh, so yeah. Okay. Great tip.
Valeriya:Yeah, that's. Okay, I tested it. And, if you have really a lot of emails, it might not be able to get over each of them, each one of them. Yeah, there are still limitations, so sometimes just play around with it. Sometimes try with the, at the beginning you have 400 actions per month. it's a lot. try to set up with every new email, see if you really need it and, then just experiment. Okay. Okay. So this is the trigger. And then you have the instruction. And instruction is basically the same instruction that you would use for custom GPT. I haven't heard of any specific rules for Zapier Central Instructions, so I just use, I just follow the same logic as if I would Be creating like a custom GPT, I would put a delimiters with something like an instruction or role and a task. And if we're talking about maybe content creation here, we do not, I would put something like an output and maybe examples, but here, for example, What's, I just set up the instructions which are relevant for me. So I told it, okay, you're my helpful email assistant and you need to check my Gmail inbox and sort out emails for me. And then I just set up like a logic. For example, I said, okay, if it's just a personal email, draft a casual, polite reply. So I don't want it to send a reply because okay, I do not recommend it either to you yet. So it drafts the reply and later I will show you my inbox as well. It looks really well and the best part for me because actually I live in Germany and some emails, some incoming emails are in German language, although the instruction is written in English, it can identify that the email is in German, and it actually drafts to reply in German language as well. I think it's really cool. and then I also asked it to analyze the contents of the email. And if it's, if it thinks the email is urgent, then it has to actually send me a slack message and add, Trello card for my Trello account so that I can see, okay, that's an urgent email. And then the second one. Currently, I exchanged the email address, but then I also set up if the email is from following senders, and then you can simply add the email addresses from maybe your business partner. So it was relevant for me. Then I asked you to add the business label to the email. And also create the draft reply and summarize it and send me a Slack message and the same also analyze it. If it's urgent, then add the Trello card that it's, that I have to reply immediately to this email. And then if the email is like marketing or promotional, just add the label archived by Zapier and archive it. And if it's a calendar event notification, then do nothing. So that's quite a simple instruction and it works well. And. Yeah, and I will show you my inbox now. And what I like about it as well is that, of course, you have to give access or you have had, you have to give access to your Gmail box, or for example, Slack or Trello, if you wanted to send you some notification. And what I really like about it is that as soon as you, connected to your Gmail, for example, you don't have to specify what exactly like to whom exactly and in which thread exactly and what should be in the body and you can specify it. And that is not only for Gmail, it's for Google Calendar and for everything else where you have the actions. But The best part is that if if you worked with Zapier automations, that you can, you have to specify everything there, or like you have to specify for automation to work. Otherwise it will be just broken. It will not work, but here it's the beauty of it. You just say, okay, have AI guess the value of this field. I said, okay, at the Trello cards and then, And then for list, for name, for description, I say, okay, have AI generate the value. So I don't even have to, think about it. I don't even have to tell it, this must be exactly this board and this, list and this description and this label. It just, Yeah, it just figures it out. I
Isar:want to pause it just for one second. First for a quick summary and second to explain what's on the screen for people just listening. So first of all, I'll start with what's on the screen. So the screen has really three different, four different components, but three that you put input to. One is what's the trigger, like what's going to start the behavior. In this particular case, an email comes in. Or like Valeria said, if you want to save time every three hour, go check my email. But it's something that triggers the process. step number two is the instructions. What do you want to do when that thing happens in this particular case, when an email comes in? And Valeria just gave it. In simple English. if it's a personal email, do this. If it's a business email, do that. If it looks urgent, do this. If it's spam, do that. Like literally what you would have told a great assistant to do as far as setting up the instructions. And then the third component is just, okay, what does. Zapier central need access to in order to do these things. So in this particular case, it needs access over to Gmail, but she says, okay, open a Trello, task or card on whatever specific board. And, If it needs to set meeting, then it needs connection to the calendar. Like whatever you want it to do, you need to be, to give it access to it. That's it. Once you give it access, it will know on its own how to do the thing. So this is like an amazing assistant that has access to everything you're willing to give it access to and can do actions literally by you explaining it in simple English.
Valeriya:Yeah, exactly. And, Of course, once you set it up, you can test this behavior. But for now, since it's running for me already for past few weeks, I'll just show you, my, my Gmail inbox. Please don't do any screenshots. so it's just basically, as you can see already from the first glance, there are these labels, right? And there's archived by Zapier on many of my like marketing emails. And I was checking it really and it is really marketing promotional emails, it never puts the archived label on some personal emails. And then, of course, I can just go to this archived ordinary, archived folder. I never do it. I just ignore them because it's, it's okay, it's not so relevant. If it's a business email, it's, and there is, for example, I was discussing the collaboration, with, with the brand And it puts the label as a business, right? And if it's an invitation from ISAR for this meeting, it doesn't put any label. It didn't create any reply because, it's an invitation to a calendar meeting. And for the reply, it's, it creates also very nice replies.
Isar:Which model is it using in the background? Do you know?
Valeriya:no, actually not. I'll have to check it. Yeah. And, yeah. And just another one, it would be like, what I was, meant to tell you is that it also replies in German language. That's where I was not expecting it, I was like, okay, it will work for some emails, maybe not for others, but then it actually created reply drafts as well for a German language. And, If you want to improve, maybe further, you could say, okay, and this is my, I can add my tone of voice and I can add a list of my emails to the prompt or maybe to the knowledge base to, to even replicate how you would reply. But honestly, I'm even happy with the replies that it creates. I rarely, Like really reformulate them.
Isar:That's first of all, it's amazing. Again, I'm blown away with this capability. a couple of questions, one for me and then some from the audience for me is there a way to give it additional data, like beyond these sources, like a, like a GPT?
Valeriya:Yes, of course. here you can connect the data sources. Data
Isar:sources. Okay, cool. And then you can put in whatever you want, like documents, files.
Valeriya:Yeah. Yeah. you can, you can even like you put, you can put, for example, the Excel here, you have the possibility or Google sheets, Google docs. I don't think you can put, yeah. Yeah. I don't think you can put PDFs, but if you have a PDF, you can as well.
Isar:Make it a Google doc and then. Make it
Valeriya:something else. Yeah. So they Promote themselves or Zapier Central promotes itself with this opportunity to connect the live data. but of course, if you need a knowledge base there, you can also put it there.
Isar:Awesome.
Valeriya:Just again, try to differentiate use cases. Like I would not, recommend you to go all in and Zapier Central and quit custom GPT because there are plenty of them. Certain use cases, for example, those Zapier bots, you cannot share it with someone. So you cannot for just custom GPTs are great as lead magnets, or if you have a team and you work together on some specific use case, maybe you have a marketing team and you need to create, I don't know, Instagram ads or headlines or whatever. And then you put all your knowledge base and you work together here. It's you cannot share it with someone. You cannot. Make it as a magnet.
Isar:Technically, it is shared because if it's a if the outcome is okay, go share this file on our Google Drive or create this thing on our Notion, then it's everybody's Notion and everybody's Google Drive. Yeah, that's true. Everybody can see it, right? There's Because it's connected to your daily tools, then the tools are shared by everybody and that solves that problem, right? it's still a very capable thing as far as sharing it. So yes, people don't have access to your GPT, but they have access to the outcome of the GPT, which is what you actually want. Zane is asking any other alternative to Zapier for creating up the AI chat spots, although Zapier is good tool, but I want to know more about alternatives being available, but definitely going to try Zapier. So do you know any other tools that do something like this?
Valeriya:that's a good question. It's, I think it's unique in its way, because it's like a chatbot for you that integrates with different apps and the live data, you cannot deploy it on the website. So it's not the chatbot in the traditional sense. You, so you cannot use it on your like lead. It's not a lead generation chatbot for chatbots. Great solutions. I know it's for example, voice flow. It's the one I've been working with. it's not just, it's not set up in stones, what you put in there, but it's actually more conversational. It can go different directions. You can also integrate it with make, for example, if you want to, I don't know, to capture lead on your website. And obviously you can deploy it somewhere else. even custom GPTs, you can deploy somewhere else. yeah. And, yeah, maybe again, it's, I think all of these tools are relevant and useful and there are very specific use cases in which one or the other one is better.
Isar:I agree. There's a question from Elsa as far as data privacy, especially when it comes to, proprietary data is involved. I will say my two cents and then you can say yours. if using Zapier anyway, you're using Zapier anyway. So if you're already a Zapier user, Zapier already had access to your email and CRM and calendar and Slack and all the other stuff that you're using within Zapier. So if you're doing that, then it's not a problem. If you're not giving it access so far, then you just have to decide what. Is where to draw the line and what's acceptable and not acceptable. And you have to do a very good job in training your employees on what's acceptable and not acceptable. So you have to have very clear guidelines on not just for Zapier, but for any of these tools. What data you, you should and shouldn't share with them. And it needs to be very well defined. Do you have anything to add?
Valeriya:Yeah. I know that's because this question was asked also to like AI director of Zapier and he said that when it comes to this conversation that happens in the chat, the models are not trained on this conversation. So that's what, that's what I know. And, yeah, regarding the automations, it's, they have information on their website. So you can, if it's really like the automation part, which concerns you, that it, it's, and then it checks the emails or you give it access. So obviously when you connect it, it asks you or it tells you on the screen. Okay, you're given the access to this app to your email account or to Google account. And then you have to read it and really accept it. And if you want to. to read about it more than you can just check out their FAQ section or website. But from what I know, at least the data in the chats, here, it's not used to train the models or for any other purposes.
Isar:Awesome. Should we jump to another use case?
Valeriya:Yeah, let's do it. So the calendar one and for calendar, I was thinking, okay, what can be like the most relevant for me? I have my Google calendar and, Yeah, I already have meetings there, and maybe it would be interesting to I don't know, to delete meetings from the chat, but then I thought, okay, I'd just rather go to Google Calendar and delete it directly, I don't need to have a chat for it, but then the cool use case I find, and I think, It would be interesting for many of you is that it actually checks and, let me show you my behavior. before the event starts, so I gave it access to my Google calendar. Again, you can, if you're using Outlook or some different, apps, you can also find them. 90 percent sure here. So before the event start, two hours before the event start, you can set up one day or one hour, everything you want. It would check my Google calendar. And before the event start, it would analyze the event description or meeting description. It was browsed the web. To return the relevant details about the event organizer and the company. For example, their LinkedIn profile or maybe recent news about the company. And then it would send me a single message in Slack, which describes what he found on the internet. And maybe if some ideas if I need to prepare something for the email based on the context and the description. And, yeah, I also said if you're allowed to also cancel or move the meeting, but obviously I would have to go in the chat to, to make it myself because that's something that must be triggered then by me. Um, and let's see, for example, let me open my, let me open my Slack. The meeting we are now, it was in my Google Calendar, and as you can see that I got the, the notification in my Google Calendar, channel, and it said, okay, you have, this is the event link, and so on, and then relevant details about the organizer and company. Yeah. Multiply AI, LinkedIn profile, recent news, Leverage AI podcast. So basically everything, like it actually gave me the, the link to multiply AI because I think it was the organizer, not ISAR, but if the organizer is ISAR. this person, it would also provide me a link. For example, it would actually go and look for a LinkedIn profile of the person and would provide me the link. So for me, for example, two hours before the event, I can check it out and I can say, okay, that's, I want to prepare for the event. And, yeah, and I can go to the website. Maybe I already have some, I know, who am I talking to and something about their company and so on. And, I want to add
Isar:something to this because I, first of all, brilliant, like absolutely brilliant, both the previous use case and this one, but my head spinning so fast right now, because the other thing that you can do that we all do, but we just do it ourselves and spend time doing this before meetings, we would go back and check our CRM. We would go to the CRM and say, Oh, what is the status with this particular client? We would go to our payment, system to see if this is a financial conversation, when was the, they pay their invoices. They didn't pay, give me a summary of that. Or we would, if you're using some kind of a note taker. So I use fathom for all my meetings, but whatever you're using like fathom, CBIL, There's so many of them now. you can go, I do that a lot. Like I would go to the actual fathom and jump to different segments of the video that I was doing. Off the previous meeting. That was four weeks ago. And I don't exactly remember everything we talked about. So I will go to see exactly what was said. I can actually bring all this information in, and I can ask you to go to my CRM and get the data from there and go to my accounting system and get the data from there and go to my note taker and read the transcription. It doesn't actually have to watch the video and get the data from there. And you can come. really prepared to every single meeting. And then the next step that, again, I'm just sorry that I'm stealing this. I'm just going crazy with ideas here, is you can actually prepare a meeting agenda based on all of that, that you can now share in an email to everybody else. Say, Hey, based on this and that, this is what we're going to talk about and send it out two hours before the meeting, without you having to spend too much time. Any second on this, or again, make it as a draft, but you can check before it does anything weird and then send it out.
Valeriya:Yeah. Yeah. Like really unlimited ideas. And, I will share the prompts, but it's the best if you really sit and think, okay, I get this information, but as I just mentioned, What can I do next? If I'm organizer, shall I then somehow summarize this information and send it to attend attendees? Or should I do something else with this? So there is really it depends on how you usually would do it, how you would approach it as a human. And then you would just try to replicate the process, in with disintegrations or with the AI.
Isar:Amazing. Okay. next use case. Again, this is absolutely phenomenal. And again, I think the interesting, the cool thing that they've done, and I started playing with it, but I didn't click until I'm seeing what you're doing, which is they basically took Zapier that we know and love for years. And a, they made it accessible to anybody. Like you said, you can literally build quote unquote zaps in here. So without knowing how to use app year and all the different definitions and how to set it up and how to select the data, like all of that literally goes over. You can create old school, traditional zaps just by explaining in English, what you want, that would happen that in some cases was even impossible. Like some things you're doing. Are not doable with zaps, but on top of that, they've added a layer of AI that actually understands the text and understands the messages and understands the context and can write stuff on your behalf and can summarize things. So the combination of these two things, being able to create zaps that were not possible before, just by explaining them and adding an AI layer on top of that is. Stupidly powerful. Like I have no good words to describe, but it's amazing.
Valeriya:Yeah. And you should not underestimate it for those who worked with Zapier. Again, that just having AI generate the value, you don't have to set up a specific, message and from where it should be sent and where exactly it's just you see purely everything here is have AI generate the value for this field and it works. I don't even have to put the effort in the first time.
Isar:There's a list of actions that it's actually doing in the background. And there's 50 lines, I don't know what they are, but like 50 lines of different things that it can do. And literally 97 percent of them just says have AI generated versus Valeria had to pick what it needs to do and it figures it out and it does a good job, which is amazing.
Valeriya:Yeah, and same as you see, of course, I provided it access to my Google calendar and set up the email, but then everything else like calendar is search term and start and so on. It's just all AI, so it just, it just works automatically again, even the first time you don't have to put that much effort to set it up. So even this. Let's say learning curve is flattening more and more so you just even you don't even have to sit one day there to think of all the different integrations and opportunities and possibilities, because it's all AI. Yeah. And let me present another use case. I think quite relevant also for many who either create content, for example, on LinkedIn, or generally, let's say if you're interested in like recent news, we are all, I'm sure you're all are interested in the news in AI space and generative AI. And for me, it's okay, should I pay for this fiddly account? Or, Should I just learn from LinkedIn feed? And maybe sometimes I myself want to post something if it's really like a hot topic. And, here, for example, so Zapier bots, they also, of course, have access to the web. And I set up a bot, which, here it works when I message it. And that's another actually cool feature in it. So you don't even have to have a specific trigger, like state. email, which, which is incoming or the meeting, you purely can use it when you message it, right? you can say, okay, it's, when I messaged the bot and you can set a specific phrase or keyword, I set news here, but you don't even have to do it. so it just, it will be triggered when I, when I send it a message and it will go through the, most relevant and recent news about some, in my case, I would be interested about AI generative AI, And it would summarize them. I asked you to make sure to include the source, the, short summary, and the link to the article. And also make sure to use reliable sources, Forbes, and maybe OpenAI blog, and New York Times, and Technology Review, and so on. And, the next step, I would also tell it, for example, to evaluate each news article, if you are creating content with it, for example. Evaluate each news article you found on their potential to be relevant to my ideal customer persona from one to five. And then if asked so, then create a compelling LinkedIn post draft with the news summary and the link to the newest source, for this most relevant or for this highly viral post, let's say. And then I would provide some context and even maybe the example of my, so my ideal customer persona as well, the description and example of my LinkedIn post where I share some kind of news or new features. So that's just something like, again, the prompt structure, it's something you would use as well. If you just prompt chat GPT or if you create custom GPT, just tasks followed by, output and followed by some context and maybe some example. And, yeah, here, so again, it's not, it does not run automatically, but here I will show you, for example, find the most recent news. I'm just typing in the chat. It should know that it's, the recent news must be about generative AI or generative AI. And, yeah, as you can see, it's, it uses, so the parser by Zapier, I did not even, yeah, okay, I think I did set it up here in the behaviors. Let me quickly check. Yeah, so in the actions, I set up the parse web page, basically, but I also didn't tell which web page, so It looks by itself, like URL to bars, it just searches through the web. So I don't have to tell, okay, you have to go to this website and that website, it just searches by itself.
Isar:So again, for those of you who are listening, what it's doing in the background, she, Valeria basically asked her to go and search the web for news. And now it's going to do that, understand what's in the web pages and create a summary in the end, based on all the different websites that it's checking. on its own, and then give her a summary of what it learned without her having to do anything other than trigger it.
Valeriya:Yeah. And also the, great part about it because, maybe for those of you who, don't see it currently, so it gets some, mistakes and maybe you saw it also from the already from newest chat GPT version. It just tries to correct them by itself. It would work the similar way if, for example, with your emails or with your Google calendar, if you actually test this behavior. So of course it works automatically somewhere in autopilot, but it would go through the similar steps if you test it. I'm not sure, but now it doesn't seem to, it will be correcting itself. Every time
Isar:you would do a live, something has not worked. that's the rules of doing a live session.
Valeriya:Yeah, but let me, because it was working before. Find the recent news. Maybe I tried to be just,
Isar:just find one that worked in the past
Valeriya:generative. And the thing about it is also, for example, Zapier central is more like a better product or they released it and they're still testing. They're looking forward to get more feedback and so on. And, yeah, and also sometimes they would change some feature or change something on the background, and then you might realize, okay, it doesn't work as it was working two weeks before, or something has changed. ideally it would work better, than before. Yeah, as you can see, I, currently it's, It searched for something and we can even check if the links are correct. So it said, okay, state of generative enterprise, Deloitte from Forbes. we can see if it's really Forbes. Yeah, it's really Forbes. Okay. It's already good. So the links are correct. And, yeah, obviously you can also say, because my prompt was actually included that I wanted to generate as well. The post potential post, and let's see if there is.
Isar:So I want to touch on this just as it's running in the background to explore, to expand on this, right? So this could be used, this kind of process for any other kind of research you want to do, right? This could be, about your competitors. This could be about what's happening in your industry. This could be about what you're happening with your. Clients industry. This could be about any kind of summary, weekly summary, daily summary, of whatever you want, but this could also be internal. Like you could connect it to an internal database of, give me a report on all the unpaid invoices in our system from the last 20 days and run it every time you want to run it. And it will give you a report. And you can even explain with format you want to report in And then you can export it as a CSV or do whatever you want with it. So the ability to create these data research analysis and reporting tools on the fly on literally any business topic you want are absolutely incredible and invaluable. there's so many things you can do with it. and it's literally just limited by your imagination or by your business needs.
Valeriya:And I just realized that I actually connected LinkedIn. And now it's even open. It's even, published on my LinkedIn Automatic. I'll just delete this one. But you saw the idea, so yeah,
Isar:so everything is possible. Those of you not seeing what's happening. There's a new post from Valeria on LinkedIn right now that was generated based on the summary from the news articles that bot just found and summarized, which is mind blowing.
Valeriya:Yeah, probably I won't be a like, I won't be doing it from my personal page. But let's say if you have a company page, or which you always constantly need to publish content for, and then you first set up this and test it, if it's more or less in the way that you would publish it, you can even set it up that it publishes it like automatically that it finds maybe recent news every day, and then it just sends out without your even involvement. And you always have, those Creators who share news, viral news, they're very popular. Or they have a lot of followers. So maybe you could do it for your company page if you are happy with the outputs.
Isar:first of all, people both on LinkedIn and on the Zoom chat are saying that your examples are amazing. The fact they're real world, really helpful examples. So kudos for selecting, those examples. Also, there's a question from Deb. She's saying, Zapier bots are triggered manually. What are the benefits over, GPTs? And I think it's very simple, right? they're connected to all these platforms. Like GPTs are not like, even if you triggered it manually, the fact that it knows how to summarize it, create a document, publish it on Slack, LinkedIn, or other stuff is just not doable from GPTs.
Valeriya:Yeah, exactly. And live data again. I didn't show you much use cases. I just maybe like really quickly. I don't want to go through it very in details, but something like, for example, I don't do it often, but sometimes I just have a data source with all the different templatized LinkedIn posts. Maybe you know how, I don't know, Justin Welch would write and then he has these templates. And then I just connected it and then maybe you would put new templates there. I would put it to my Excel sheet. Then I, here, I don't have to update it. And then I also set up a prompt, which is okay, I would put something like a topic there and, just find a suitable template and generate me, from the template. And then it actually tells me usually, yeah, okay. I will use template two or I will use template whatever then, and then it generates the kind of post for me. So this live data sources, because. Again, in custom GPT, you would have to, if you have new data, updated data, you would have to go back there and delete the data source and upload a new one. And it's okay, it's not very handy. Or sometimes there are use cases for custom GPTs. everyone was talking about, okay, you have, you can do like an onboarding bot. But what's the point of this onboarding or training bot for employees if all these onboarding instructions, they change every week, right? Then you'll have to just, you'll have more to, to have more work to update them constantly. But here you just connect the notion database. If you have all your instructions, Like standard operating procedures or something like this in Notion because they would be updated obviously, and then it's updated. It's updated automatically. And then you just trigger it and ask the questions about always like a new document.
Isar:Amazing. there's a question. Can you configure Zapier? A question from Albert on Zoom. Can you configure Zapier? So it only triggers when a new email arrives meeting a specific requirement. So versus every email. So Zapier, again, I'll, I'll let you answer.
Valeriya:Yes, sure. And, Here, because I didn't want to share with you, or I didn't want to show you the actual emails of people who are relevant, like for my business, let's say, but that's what I did previously. I actually tell it to, okay, if the email from the following senders, senders, and then I like edit 10 to maybe 20 emails, then at the label business to them. And currently I just replaced with my like other email. You can do it, and it understands, and it really then says, okay, it labels my emails, then okay, it's a business email, for when it comes from these people, or otherwise you can also set up, also read through the email, and if it's a business context, then, do this, or again, you can just specifically tell it from which emails, the, they must be labeled as urgent or as, I don't know, a reply right away and sent to Trello and so on.
Isar:I think the question is, was to trigger Zapier. can I have it not even triggered unless the email has this or that on the header? Yes,
Valeriya:that's, you can do too. And,
Isar:like in Zapier in general, those of you who don't know, Zapier has a pretty good filtering system. And so there's like a Zapier filter function within Zapier that allows you to, and you can probably use the same thing here, right?
Valeriya:Yeah, you have you should have it because here, even like you see, if you edit the trigger, you just have to go back to the trigger and see, for example, currently it's an inbox. So every email that I receive, it would be triggered, but there's important, for example. and you can set up and spam and specific categories. So for example, if you add in your inbox, this specific category, which obviously you can do, then it will only work for the specific category, this trigger.
Isar:Amazing. By the way, now, something interesting for those of you who are a little more advanced. Let's say you don't necessarily know in advance, you don't have an automation in your email to put the right emails in the right categories. You can use one Zapier to put the things in the category and then use a different one to just do different things on specifically these categories. So there's again, the. Once you understand how this tool works, there's literally no limit. Like you can use step one to do one thing and then step two to then only apply for what step one did. there's a bunch of additional questions. so Duane is saying, I just set one up, for me while I'm on vacation next week, and told it that if the email has more than just me on it, then ignore it because then there's other people, but then if not, then respond in different ways. Again, there's. Multiple, multiple ways to use this. but Maria, this was absolutely incredible. Again, you, first of all, the way you shared it was very detailed and very thorough, and also the use cases that you picked were fantastic because as you mentioned, they're applicable to literally anyone, if people want to follow you, work with you, get your content, what are the best ways to do that?
Valeriya:Yeah, just follow me on LinkedIn.
Isar:If you find her, it's her.
Valeriya:Thank you so much. It was a pleasure.
Isar:Thanks everyone. We'll see you, next Thursday. And if you want to check out the course, don't forget to do that. And until next time, have an amazing day, everyone. Bye bye
Valeriya:bye bye.