Leveraging AI

198 | Build a Business-Boosting AI Agent That Books Calls While You Sleep-No Code w/ Carolina Posma

• Isar Meitis • Season 1 • Episode 198

Share episode

🎉 Special Announcement: EPISODE 200 is coming June 17!

We’re going live at noon ET with The Ultimate AI Showdown — 4 of the world’s top AI experts, 4 tool categories, and everything you need to know to pick the right solution for your team.

👉 Reserve your spot here: https://www.linkedin.com/events/theultimateaishowdown7338224988014977026/theater/ 

👉 Fill out the listener survey - https://services.multiplai.ai/lai-survey
👉 Learn more about the AI Business Transformation Course starting May 12 — spots are limited - http://multiplai.ai/ai-course/ 


You don’t need a tech team to make your business smarter. In fact, after this webinar, you might not need one at all.

Join us for a step-by-step session with AI strategist Carolina Posma, where we’ll show you how to build and deploy an AI agent that are trained on your company’s content, communicates with clients and prospects through DMs, and even books sales calls directly into your calendar.

Carolina has helped everyone from Instagram influencers to major e-bike manufacturers transform their customer interactions using FlowGent.ai — a platform that makes AI agents accessible to anyone, no code required. In this session, she’ll break down her process, show a live use case, and teach you how to do it yourself.

Carolina isn’t just an AI builder - she’s an advisor to Flowgenz and an in-demand corporate AI trainer across the UAE. Her focus? Making sure businesses don’t just “experiment” with AI - they deploy it effectively. She’s hands-on, deeply strategic, and knows how to turn theory into ROI.

About Leveraging AI

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!

Isar Meitis:

Hello, and welcome to another live episode of the Leveraging AI Podcast, a podcast that shares practical, ethical ways to improve efficiency, grow your business, and advance your career. This is Isar Mateis, your host, and we have an amazing session for you today. We've been talking a lot about agents, both in previous episodes as well as on the news, like agents are now, everybody's interested in agents, everybody wants to build agents. Uh, everybody wants to know what agents can do for them. And the reality is most people don't have a clue how to do that, how to get started, which platforms to use, and so on. And this is exactly what we're going to tackle today and we're gonna demystify the whole concept of creating agents. We're gonna show you exactly how to build an agent that you can actually use that will be very useful for probably most companies in the world today, big or small. So what are we going to build? We're going to build an agent that will learn everything about your company just by looking at your website, and we'll be able to answer questions on that website to, well, wherever you wanna place it for a demo today, we're gonna place that on Instagram. So we're gonna build an agent that will be able to answer DMS on Instagram. Based on information from the website, you obviously can build it and connect it to any other thing like place it on your own website and so on. And the cool thing is you'll see that the Egen universe, that while this sounds complicated, it actually really easy and will do all of that within the next 30 to 40 minutes, which is incredible when you, again, most people think that it's really hard and really complicated, and you gotta have an army of coders to do this. The reality is it's not, and we're gonna do this with the help of Carolina psma, who is our guest today, she is an expert on doing this. She has her own agency helping companies do exactly that. She is a. Advisor, like an external, uh, promoter of Flo Gent, which is one of the most commonly used platforms today to develop agents. And this is the platform we're going to use today to show that to you. Her background is in marketing and in entrepreneurship. She's been in multiple marketing roles across multiple, uh, companies, including several of her own. So she understands business, she understands entrepreneurship, and she understands agents, which makes her the perfect guest to share with us, uh, her information about this. So I'm really excited about this one. Carolina, welcome to leveraging ai.

Carolina Posma:

Thank you so much, ISAR, and thank you for this, this very kind introduction. So what I want to show to you today is indeed how you can build an easy and Instagram ai DM appointment set. So you already mentioned it in your description, ISAR, uh, an AI agent that does sales for you on pil, right? So that's what we're going to explore today. So indeed, uh, pick about me. Uh, oh, sorry. Yes.

Isar Meitis:

No, I'm saying that's, that's amazing. I think it's very exciting. I just wanna, first of all, thank everybody in the audience, the people are joining us live, whether on LinkedIn and or on, uh, zoom. So thank you so much for being here. If you have, uh, first of all, go ahead and introduce yourself. So where you are, what your experience with agents, so we have a little idea of who's in the audience. And, uh, if you're not in the audience, then know that we do this every single Thursday at noon pm Eastern. You can come and join us, and then you can ask questions, you can chat with the people in the chat. You can network and so on. Also, I wanna mention one last thing. If you wanna learn more in depth on how to leverage AI in your business. There are two ways we can help you with that. One is we have the AI Business Transformation course, which we have been teaching for over two years. The next session opens on the second week of August. It's a four week course, two hours a week. That will give you an incredible level of understanding in AI and how to apply it in your business. Whether you're just a business person or you're in leadership position in a company. Uh, in both cases, it will give you incredible. Boost to your understanding of ai. The other thing is we work with companies on custom tailored workshops to help businesses accelerate very quickly into the AI era. We do this either in person or online, depending on the setup of the company. So if you're interested in any of those, just look for the links in the show notes, uh, or just connect with me on LinkedIn and ask me whatever you wanna ask me. Uh, and now let's dive into agents and let Carolina lead the conversation. Uh, Carolina, it's your stage from now.

Carolina Posma:

Yes. Thank you so much. So let's talk a little bit first about the difference between chat GPT, AI automation and AI agents. If you are a frequent listener, uh, you might already, uh, know the difference, but I always like to, uh, go back to basics to just, uh, get everyone on the same page. So I will share a couple of slides first, where we will discuss this. So let'S get started. First a little bit on the difference between chatt AI automations and AI agents, because today we're going to talk about AI agents and we're going to create our own AI agents as well, but. As I said before, it's always very useful to get a little bit back to basics on what are we actually talking about. So first, where do we stand Now in the world of ai, many people have already probably seen this, uh, lead note from the Shopify CEO telling us that, um, uh, telling all his employees that before asking for more headcount and resources, teams must demonstrate why they cannot get what they want done using ai. So this already emphasizes the importance that huge companies lay on understanding AI and using AI and knowing how to get things done using autonomous AI agents, as it's mentioned over here. Then we have companies such as Gumbo that are only with four persons that are raising 17 million, uh, in dollars. So they're building, as they claim, the first 10 mil, uh, 10 billion, um, uh, 10 per 10 person, 1 billion company. We have companies looking for AI first talent only, and you have to automate your job as you go. And even further, there are companies like Fire Crawl who are currently hiring AI agents. So they're not even, uh, looking for people to build AI agents. No. They have job agencies for AI agents open. And the great news here is that we are still in the very much beginning and the people who adopt this technology right now will win just like in a previous stance like internet, smartphone, those are very disruptive, just like AI are. And. When we're talking about AI agents, the first thing that you need to grasp a little bit is Chat g pt. But I am assuming that everyone right here has played around with chat. GT already has had a conversation already with chat, GPT, uh, but a lot of people are, uh, are calling chat GT and AI assistant. And this is where I want to get started because in my opinion, AI uh, chat g PT is just an AI advisor because chat GT is brilliant to get advice to brainstorm ideas, to create content that will help you when you're stuck in anything. Chat GT is your personal advice that you can ask anything, but one of the problems of chat G PT is it can't really do anything. Yes, if you're very advanced with Che gt, you may be aware of custom Gcpt, but in general, Che GT doesn't do, uh, anything. Then we get on to automations. And automations was already possible for a very long time. Uh, let's take this, uh, super straightforward automation example where you're connecting two different tools. Let's say you have a, uh, lead form on your website and people can leave their, uh, information over their first name, last name, and you want to have that added to, for example, your CRM such as HubSpots pipe drive for any other CRM. You can do that using tools like Zapier make and it m uh, tools that were there for years already. So putting this information from one side to the other side was already possible for years, years, years. But why is there all the hypes suddenly about AI automations or AI automations? So process automations as we know them, completely changed with the introduction of better cost friendly and accessible ai. 2022, the, uh, A-P-A-P-I of GPT-3 0.5 became available and suddenly. AI wasn't anymore just for the big corporates with huge budget. No, it was for all of us, for you, me, everyone who, uh, is, um, yeah, open to play around with ai, even as an individual. And what we were seeing was that AI was getting so much more smarter and at the same time, so much more cheaper. So it became more, more, more available. And what does AI now allows us to do, to do in these process automations? Well, it allows us to interpret information, to summarize information, to turn unstructured information into structured data. So what do I mean by that? Let's say for example, that you have, uh, uh, someone sending in a resume or a cfi, it can be sent in a PDF and that's a huge bulk of unstructured information. It can, the, the name can be anywhere on that PDF, right? But with AI, it can read that information and it can. Take out, for example, oh, this is the first name. This is the last name. It can turn unstructured information into structured data that's so powerful. It can create new content. So there's so much that you can do with AI right now. And when we take a look at our example, and please interrupt me if you have any questions, but if we are looking at this example from, uh, the lead form to, uh, name, email, website to your HubSpot pipeline, suddenly you could start doing stuff with that information. So with the introduction of ai, you can can summarize the website, for example, you can um, give some qualification criteria to AI and ju have a judge if this is a qualified lead or perhaps you have a sales team that's pick up the phone for everyone who is, uh, who is filling in that lead form and. AI or an AI can help you create a call script for, for that salesperson, right? So there's so much that you can do with ai. You can interpret information, unstructured information into structured data. It's changed the rules. So right now we discussed chat, GPT, which is your personal advisor, to ask questions, get help. Then we have automations. And the old school automations help you to stop copy pasting info and doing the same boring tasks. It helps you to connect different softwares and tools with each other. And then since a couple of years, we have AI automations and that's making your automation smarter with ai, summarizing it, interpreting it. But something new in the last couple of months, year, uh, happened and that that is AI agents, and this is even a step further. So let's move on to the AI agents. You can think of AI agents as Jet GPT, but it doesn't just chat, it takes action. So you can train an AI agent, for example, in your company's knowledge, and you can plug it into different tools. It can book meetings, it can update your end records. It uh, records. It can send messages. It gets real work done. Before we dive into AI agents and how to build, uh, it's without, if you don't know, have any technology at all because, uh, like is, I said, you don't need to, uh, to be able to code. No, you can really, there are softwares that allow you to just look and play with AI agents. There's something that you first need to know it. Where we stand with AI agents right now, and what I'm showing you on the screen right now is the triangle of talents. And, uh, in this triangle of talents, you see a triangle. It looks a little bit like the social hi hierarchy of different levels that employees can have. And level one is lowest level. That's useless. And that's someone, if you tell them what to do, they rarely do it right. If you go one level up, we're talking about task conies for those, uh, employees, persons should tell them what to do and how to do it and when to do it, and it'll get done. Then we have a third level of talent, which is problem solver. Tell them what to do, and they figure out the how. The fourth level of, uh, talent is systems thinker. Tell them. Um, and I do see that the credits actually are missing. This is, uh, shared by shampoo on, uh, on, uh, Twitter. The fourth level is systems thinker. Um, tell them the problem. They set a, a, a system to figure it out. People plus process and level five is a superstar. They identify the right problem and get it solved. So where are AI agents? I wanna pause

Isar Meitis:

you. I wanna pause you just for one second to add two things, one to do a quick summary of the great intro you did of the differences between AI and agents and automations, and I think most people are still struggling with that, so I think that was great. The biggest differences between agents and just ai, large language models is, like I said, they can take action. The second thing is they can figure out their own steps. Like if you wanna do a multi-step process with, uh, ai, large language model, you gotta tell it what the steps that it needs to do. Even though the reasoning models solved it partially, but only partially, uh, and the agents know kind of how to think for themselves. You give them the goal, they figure out the steps, versus you have to take them the steps. And number three is they know how to work collaboratively with other agents. You can build a hierarchy of agents that can talk to one another. When it's LLM, it's just that one LLM. So this is kinda like a quick summary and my 2 cents to add to that on what you just said about the, the talent, there was a very interesting, uh, interview with Sam Altman this past week at the, um, keynote of Snowflake, the big data company. And when he was asked about what he's excited about, that's probably coming in the next year. So the next 12 months, he basically thinks that the biggest thing is that AI will be able to solve really big, really complex business problems that we are either unable to, or it takes us a huge amount of effort to solve just by giving them more compute. So he's basically saying what you're talking about right now, that we are gonna go from level two to level four, maybe five, uh, by just giving it access to the right data and just paying for a lot more tokens. But then it will know what to do. And for, if you can solve a really big problem for a really big company that might be worth tens of millions, hundreds of millions, billions, depending on how big the company is. So saying, okay, I'm gonna throw$2 million worth of compute at this. Makes perfect sense in those scenarios. And so just to give you a broader idea of where we are, where we are going, uh, this is kind of where we are now. I'll let you continue on where we, uh, where we are right now.

Carolina Posma:

That's amazing. That's, I, I don't know. It's, it's, it's sometimes, so it is honestly also a little bit scary because even AI agent already can go to level four, the systems think go level five, a superstar route. That's, I, I think that's, that's brilliant when we get there. But also, I'm so curious on how the world will look like by then on, on, on what will change. Because so much is going to change, but at the moment, um, I think, um, right now, uh, AI agents are still, most of them are still task monkeys. I do know that, uh, some companies are playing around with problem solvers already, but, uh, the most AI agents that, uh, you for example, would create, uh, in, in softwares that you see, uh, often around, like for example, N eight N are still task monkeys. You need to tell them what to do, how to do it, and when to do it. And then your AI agents will get it done every single time, if you instructed correct. So with that in mind, um. My experience, there are two big use cases for AI agents right now. Um, AI agents can really well connect with your leads or customers, uh, via you often hear voice, uh, so via calling, via messages, for example, on Instagram like we're discussing today, but also in website chats or in, um, uh, on WhatsApp, for example. So, uh, AI agents, uh, are great right now to really get the conversation going with customers and leads like a sales employee or like a customer support employee. That's the one use case in the other use case is a personal company assistance, um, doing different stuff in the backgrounds, uh, more like an Automation plus, for example, doing some research for you every time a lead is coming in or, um, the, uh, let's say, uh, creating content for you doing research on that. So that is another big use case. If you are not technical at all, um, or at least you consider yourself to be a non-technical person, there are three software tools that I can recommend you to get started with AI agents and these software tools are for your personal company. Assisted are Zapier. They, I think they have a really good super user and beginner friendly, um, personal company, assistant AI agent builder, uh, relay app. I think, uh, I really like them in the sense that they have a lot of pre-built, uh, AI agents already for you. So you can just click and play also for the personal company assistant. And if you are looking for an AI agent builder that allows you to create AI agents that can connect with your leads on, on Instagram, WhatsApp, on your website, then I highly recommend Flow Gen, uh, ai. Right. Also involved as an at Pfizer. So let's take you look at some examples that we can create, uh, using, uh, an AI agent that can connect with your leads or customers. So let's, like I mentioned, the lead generation on your website. Uh, Instagram BM AI agent that also, for example, customer support for e-commerce stores. Uh, connect your AI agent with your Shopify and have it search for products and have it, uh, updates, uh, order status or request order status when people are asking for it. Or for example, an AI coach trains on, on, on your course. If you are, uh, a consultant or if you're running a course, uh, imagine that people can chat with your AI coach version, uh, on WhatsApp or in Slack, for example, and you, uh, giving your students or clients 24 7 advice. Or let's say for example. Uh, customer support or sales on, on WhatsApp for events. I think if you're running a huge conference, maybe you can start, uh, reaching out to everyone on WhatsApp. It's the new, uh, the newest updates and, um, if people are, have questions about where do I park or who are the keynote speakers, or anything like that, and AI agents can answer that easily right now. So just to recap what I, um, showed you here and what I discussed here. We went from kechi pt, which is a personal advisor to automations, to AI automations, and we did discussed two different types of AI agents, the sales or customer support employee or the personal assistance. Any questions so far?

Isar Meitis:

No, I think this is fantastic. I don't see any questions in the audience, people saying that it's interesting and that it's awesome, so I guess you're on the right track. Brilliant.

Carolina Posma:

No, that's good to hear. I see that NADA has a second brain assistant built on nat com. That's amazing. Okay, let's move on then to actually creating our AI agent. I am going to hide, let me see if I can hide this. Voting yes, because then I can actually see what I'm doing. So if you want to, uh, walk through this, uh, together with me, you are welcome to do it because, uh, well, I'm going to build this live, always nervous for the live demo, so I hope everything goes well. But, uh, I will show you how to do it. So, um, where we're going to build our, what we're going to build is an Instagram AI agent that can answer any questions on your Instagram dm. So let's say that you have a lot of followers on Instagram, uh, as a company or you're an influencer, then this is a brilliant use case, uh, to try and add an AI agent to it. So, where we're going to build this is flow Gen ai. This like flow and agent combined, but then without the a middle flow gen. And on the flow website, you can click on get started for free and then you can add your email address or click on continue with Google. And then you will get on a page that is the onboarding form. Like every software, uh, there's an onboarding form and there you can tell. What should we call you? Well tag Lina. Um, we're going to build an AI agents for Multiply today. Then, uh, we'll just keep this on one to 10 and we will continue. So next, uh, where did you find us? Uh, referral. What's the use case? Uh, Instagram AI agents, but you can put in any use cases will of course help the team of flown to, uh. Uh, builds their software better for you. So, and this is where the magic is going to happen, because what's going to happen right now is, uh, if you fill in your website over here, and we're going to do this with website, if you fill in your website over here in the background and ai, uh, uh, or the website will be scraped and an AI agent will be trained, so to say. So, uh, it'll be added as knowledge and a prompt will be generated, uh, on your website. So you don't need to, uh, have any tech skills for this. You can just fill in your website and automatically, uh, it'll be threats. So we're building your agent now, and in the same time, you are going to be let, uh, guided through the apps. I'll just quickly go through that. So, uh, the credits, uh, dashboards, um, if you are doing this on your, uh, on your own, you can easily read it, but they have a chat manager. Uh, you can have different agents. You can, uh, connect with different integrations and settings. Uh, now click on agents and click on agents overview. Then you can open it to see your first agent, but the agent is at the moment, uh, still being built. Uh, this is your agent's playground. So once the agent is ready, we can play around with the agent and we can chat with it. In the meantime, let's take a look. You can edit to your website here to edit your agent, and we will take a look at how to edit your agent. So what we will see in, uh, a couple of minutes is the agents, uh, the system prompts being, um, being filled in. So. For every agent, the system prompt are the instructions that you can explain to the agent on how to behave. So you can really think of it as, um, the introduction or the onboarding guide for any of your employees. In your onboarding guide, you also tell, um, how, uh, someone should behave or, uh, what someone should do, or, uh, what's expected in what cases. And that's the same, uh, uh, for the system prompt. So. Remember that the system prompt is really the heart and the brains as well of your AI agent because you are, uh, telling, its, its personality. You're telling it, um, the way to behave, which actions it you take, um, and all in plain English. So this is really the same as as chatting with chat GPT, but then, um, you're, yeah, explaining it to your AI agent on how to behave. The next part of every AI agent is that an AI agent should have before, before you

Isar Meitis:

continue just to talk about the system prompt. So, and I think you framed it perfectly, right? It's, it's the onboarding and the training of a new employee, right? But it covers, I. More than that. Just like you said, because it also covers personality, it also covers how you're gonna answer it. Do you want to be brief? Do you wanna be detailed? If you are detailed, how detailed do you want to be? What's the format of your answers? Like every one of those things you can customize and define for the agent exactly how to behave and different than the average employee. It's actually gonna follow these instructions, uh, exactly how you tell them. So if you, my question to you though, Carolina, is, does flow gent or do you have a different resource of all the different topics that should be in a system prompt? Like what, in order to make sure you're not missing anything important, do they guide you through these different things or do you look for that on YouTube and start there?

Carolina Posma:

Um, you will see it in a couple of minutes. Okay. Uh, but, um, Flo Gent right now, uh, because I filled in the website, Flo Gent has the, um, system prompt to created automatically with ai. So they have a team of AI agents in the background that go to your website and create a system prompt automatically. So, uh, the rough draft of the system prompt is already there, uh, which makes it more easy to get started with. Uh, from there on, uh, you can make any changes. Um, things that I would include and, uh, always, and this is something that uh, yes, that I'm going to share, uh, now as well is, um, always defined the role of the, of the AI agents. So who is he or she or it's, or, I don't know how to define it, but, uh, who are they? So is it, um, uh, is it. A personal assistant, is it an, um, a sales employee? Is it an, uh, appointment setter? Is it a customer support employee? This already sets the stage first for your AI agent on how it should behave in, in, in the rest of, of the conversation. So the role then always define indeed. You mentioned it already as well, the goal. So what's the goal of the AI agent for a lead generation AI agent? You wanted, uh, maybe to collect email addresses or for a customer support AI agent, the goal is to help the customer as best as possible to solve its problems. Then I always recommend to include, uh, the standard operating procedure, but you can name it a different ways. You can name it, for example, you can say it's the task or, um, the standard operating procedure or the workflow or anything like that. But there you define usually on, um, bot steps. It should take. If you wanted to take certain steps. So for example, um, for a generation AI agents, sometimes you don't want it to, to have certain steps and you just want to go with the flow and really follow the conversation. But sometimes you want to AI agent to first ask a couple of questions that are, for example, qualifying, and then, uh, go with flow. And that's all stuff that you can also, uh, define as well. Um, tone of voice, um, examples. So this is a little hack that I, uh, always use, but. I love to, uh, add examples to, to, to the system from, because it makes it very clear for the AI agent how to, uh, behave. But what I usually do for that is, uh, screenshot a couple of old conversations that are already done by, by human, um, and then put that in chat g pt, ask to transcribe those screenshots and then, uh, put it in the, uh, agent's user, uh, set or uh, user structure. And then you already have some conversations that you can add as an example as well to the AI agent. And what AI of course, does is, is the whole goal of an LM or AI is to predict your dream outcome based on the information that you are providing. And that's also why you will notice that AI is always. Very much a pleaser because it's trying to predict your dream outcome. So the more information you can give it, the better it can predict your dream outcome. Because the more context it has, the better you can predict, uh, the dream outcome. And the last, um, uh, the last element that I would always include are rules. And rules are basically, uh, maybe you, uh, will recognize this as well is are creating AI agents. Uh, you will notice that it does some stuff that you don't want it to do, so that I just put on rules or notes or something like that. Never do this, always do that, never do that. And just when I'm playing around with it, when I'm testing it, then grills are just a bunch of random stuff usually that I add when, um. Um, yeah, when I notice that it's doing certain behaviors that I don't want it to.

Isar Meitis:

Yeah. A a a quick example of something that happens to me all the time, uh, is these tools like to start with a prefix to the answer. So it will say, oh, based on the information that you provide with me and based on, uh, the question that you asked, and based on 1, 2, 3, 4, here's the answer. I said, no, I just give the answer. And because otherwise you will say that multiple times in the conversation, which is really annoying. So basically telling it not to do any prefix and just provide the answer, I. Is something that I found very helpful. So that's just an example so you understand in your head, uh, what Carolina is talking about. But, but there's a lot of these small little tweaks. Some of them are consistent, meaning once you learn them, like the example I just gave you, uh, by the way, that's the same thing in, in NA 10 automations or make automations, right? It will sometimes give you more information what you actually need in the output, and then it messes up the next steps of the automation. So you need to very clearly define what to include, what not to include, what format to include it in. And these just are part of the rules that Carolina is talking about.

Carolina Posma:

Yes. Yeah, absolutely. When you're mentioning this, uh, it reminds me one of the favorite one is of course the M dash that's that everyone is now so afraid of, of using because it's, it's clearly signals ai. So I always try to, to tell the AI agent not to use an an M dash, you know, like the, the, the, the long dash. And its answers, it's, it's mostly not listening to me. I do try to emotionally manipulate the AI 80 by saying that if it uses the M dash, that will, uh, get fired. Sometimes it still decides to ignore me on that, but, but this is also a, yeah. Really famous, uh, example of one of the notes that you want to exclude.

Isar Meitis:

There's an interesting question before we continue. An interesting question in the chat. What if there's no personal or company website, but, uh, so that's the question from the chat, so thank you, Netta. But the, I will broaden that. Let's say I have a website, which is most companies, but I wanna add more information to the agent to have access to more stuff. Uh, how can I get, how can I connect it to additional data sources?

Carolina Posma:

Well, that's indeed the next step where we, uh, are going heading now because Oh, perfect question. You can that in. Yeah. This is actually the perfect question because that's something that you can do in the knowledge base. And what I actually want to emphasize over here is, um, that these principles, um, are actually the same for all the different AI agent builder, or not all, but most AI agent builder tools. So everything that I'm mentioning here, you will indeed recognize this. You mentioned as well in n it, n in make. Um, so the principles are the same, um, and the elements of an AI agent are always the same. Uh, the difference here within Flow Gent is that, um, it just makes it, uh, they make it very easy to connect it to different tools like Instagram, uh, and WhatsApp. And they have a chat manager and they have certain features. Uh, for example, for Instagram, uh, something like, for example, if someone comes to keywords, then you can automatically send a DM already to them, or, um, so there is a lot of prebuilt, uh, features in there that make it, uh, very easy. Uh, but the elements are, uh, very much the same. So knowledge base, um, here you will find, uh, options to expand your agent's knowledge. So right now in the backgrounds, um, the AI agent team of flown already, uh, edits the website as knowledge. So all the content that is on the website, it created an FAQ based on the website. But there's more things that you can add. For example, if you are using Notion, um, you can connect your notion and then automatically it's real-time things, the notion page that you want the AI agent to have access to. So this is, uh, one of the features that I very much, uh, that I find very useful because, um, this way you can make the AI agent's knowledge very dynamic. If you make any changes on the Notion pages, the knowledge of the AI agent will also be automatically updated. So, um, this's something that's, that I like. So connecting notion, but you can also add, for example, PDF, uh, you can just copy paste text, you can add more q and a, more websites, um, uh, pricing plans. So these are, and they are, uh, adding more and more, uh, as we go. The, the next step is to choose to pick your channel. So these are the AI agents that are really meant to, um, uh, be able to connect with customers and leads. So over here, you can, uh, edit as a chat bot to your website, just like the one that we're seeing over here. But you can also add it to Slack, uh, to Slack, to WhatsApp, to Instagram. And if you are more technical, you can trigger it with a web hook or, um, access it through the flow Gen API or have it scheduled. But, um, yeah, the most commonly used are Instagram, WhatsApp, slack, and chat bots. Then of course, one of the, and chat means

Isar Meitis:

it's gonna be like a piece of code I can put on my website. That's what it means.

Carolina Posma:

Yes. Yeah, you, you can copy paste the codes and then, um, edit to your website. So, um, yeah, it's, it's, I will show you, um, as well, but it's an, uh, yeah, it's just copy paste a small piece of codes and, um, one of the main elements of an AI agent is that can take action. So, um, you mentioned already, uh, before as well, is our, an AI agent should be able to make decisions. It can, can, um, yeah, make decisions based on the goal. Uh, it should have knowledge. And one of the big things of an AI agent, which really sets it apart from just general chat bots or chat assistance, is that an AI agent can take actions. So it can connect with your software, it can connect with your tools, and it can take actions in those software and, uh, those tools. So in, uh, flow gen, they have, uh, a couple of, um, pre-built actions already, such as sending a Slack message, um, car.com. I think they will release that next week. But, uh, booking an appointment, kindly booking an appointment, uh, sending email notifications. But over here, um, they have a custom tool builder, which I will also show where, which is really, uh, comparable to make.com. You can build multi-step, um, tools. You, uh, can connect it with any API. So this is really, um, how you can make your AI agents so, so advanced. So what are things that you need to, to be thinking of? Uh, for example, I helped a company that wants the AI agent to process and create invoices and process that invoice information. So I, um, created a tool that's, uh, connected to their invoicing software, and then at the AI agent could decide when it was necessary to actually create an invoice. And then at those moments, it creates an invoice. Other things that you can think of is, for example, adding, having the AI agent at leads to your CRM. So to connect it with HubSpots power drive. Um, other things that you can think of, uh, is, uh, adding todos, for example, to your project management tool connected with Asana, connected with notion mandated com, any project management tool to f the AI agent add to. When necessary in a way that you can have this AI agent actually decide on, um, on when to do it, is by in the prompt. So, um, like I showed you before, the system prompts, um, you need to first connect the tools and then in the prompt you can tell it, use this tool when you need to create an invoice or use this tool when you want to book an appointment. So this really, you, yeah, you really tell the AI agent, and this reminds me that this is one of the elements that you should also, uh, add in a prompt is the tools that an AI agent can use and when, so the process from

Isar Meitis:

that perspective, again, just to clarify to people, is you need to know, and it's the same by the way, in any agent creation platform, whether you're creating this. On flow gent or any other platform, you need to define and create tools first. And tools are sometimes the tools themselves and sometimes tools that you make up, right? So a tool could be connection to Slack, okay? Slack is a tool, but sometimes it could be a tool that you define that is something that the AI can do. So a tool is basically, think about it as something the AI can now do and you can explain to it when to use this thing in the system problem. So this is, again, going back to the onboarding, it's the same exact thing. If you are a new sales employee, uh, you got a new lead, go and log it in the CRM, here's how you do this. So the, here's how you do this, this is how you build when you build the tool. And then just explaining that that's what needs to happen, happens in the system. Prompt.

Carolina Posma:

Yes. Yeah, exactly. So, um, yes, and the last thing that you want to decide, of course, is what kind of AI model do you want your AI agent to use? So right now, uh, if you're on the free version, uh, and I'm showing you the free version right now to show you that you can just try everything for free. You can only choose between four oh mini and 4.1. Um, I always recommend in that case to use 4.1. And in general, I like, uh, right now, as it's now June 12th, maybe Jet GPT will, uh, uh, uh, release GPT five soon. But right now, I, I would recommend, uh, using GPT-4 0.1, but if you, um, upgrade to a page plan, you can also use clouds. Um. They're adding, uh, O three, uh, GT O three as well. So if you're not familiar with the different models, um, 4.1 is, uh, a model that you can't see in check GPT, but it's available for, um, yeah, the API, which means that you can, uh, include it in, in automations like these. Um, I personally prefer 4.1 over, uh, four oh. Uh, but when you're creating your AI agent, just try the different models, see how different your agent is responding. Um, once you pick your model and create optimizer pro after that, then when it's live, don't randomly change the, the, the model again later on because the agent can start behaving differently, even if, if, if it's the same prompt when you change the model. So that's something to be aware of

Isar Meitis:

a hundred percent. What it basically means is for different use cases, different models are gonna work better and your only way to know, sadly, is to actually test it out yourself.

Carolina Posma:

Yeah. Sometimes, um. For me, I do notice that 4.1 is really the, it does it all model, but before we had 4.1, you could basically choose, choose between 4.0 and or cloud 3.7, and I would always go with cloud 3.7, um, because it's, uh, it's, if it's was more complex AI agent where needed to make, uh, different tool, uh, different choices between different actions. So in my experience, cloud 3.7 was really, uh, good for an AI agent to actually make those choices and know when to use the tools. And GT four O sometimes forgot to use the tool, kind of the was, but GT four, four O was again, better for, for conversational experience, was a little bit more human, I would say.

Isar Meitis:

Interesting question from the chat, uh, from Gwen. So thank you Gwen, from the question, what about data security and. Privacy, right? So you're gonna give it access to your company website, you're gonna upload documents, you're gonna give it access as you to your Instagram account or your Slack account. So how do they maintain, where do the data go? Uh, what risks are company taking by integrating stuff like that?

Carolina Posma:

Yes, that's a good question. So, um, I know that they store all information on, uh, in the eu. Um, the software is GDPR proof. So, um, and with the open AI models, uh, that's also good to know that, uh, open ai, if you are using the API of open ai, which is being used there, um, GPT, uh, or OpenAI isn't, is at least claiming that they are not training on your data.

Isar Meitis:

Yes. Just like

Carolina Posma:

Google is claiming things and you, you, you can never be home. Sure. But they at least promise that they're not training on your data. Okay, then, uh, yes. So then we have, uh, we end the onboarding, uh, sequence here in the playground and in the backgrounds, the multiply AI assistance is already trained. So if you try this, uh, yourself, you will see, uh, also an AI agent trained for you. Right now, this agent isn't connected to any tools yet, but you can chat with it already. So, for example, what services, uh, do you offer? It should also be, it's, I think these are the colors already. Yes. It's this light blue color from the website. Um, and here we see that it's using the terrible M dash, but it's, uh, answering correctly, hopefully. So we offer a full range of, uh, AI powered solutions. Um, maybe we can ask, um, something like,

Isar Meitis:

you know what I have, I have an interesting question to test it with. Let's try that. Yes.

Carolina Posma:

I love that.

Isar Meitis:

Uh. Which podcast episodes were about AI agents? Let's see, because, because I have, we have all the episodes there transcribed and let's see if it knows how to go through that data as well. That will be a very tricky question.

Carolina Posma:

Yes.

Isar Meitis:

Uh, because it has hundreds of pages. So what did it say? Can

Carolina Posma:

you tell me it? Okay, so it's search Knowledge. It's uh, we have great episodes. So this is the Powerful AI Agents and it's is, oh, I'm, I should have opted in nta. It is also sharing the link over here. Very cool. Yes. But we can maybe ask for more. Uh, can you give me three episodes of the podcast, uh, that were about AI agents? Let's see if it can go even further. Even more.

Isar Meitis:

Yeah.

Carolina Posma:

Search knowledge. Absolutely. Here are three episodes. Let's see if these are indeed about ai. H you know this better than I do.

Isar Meitis:

No, this is really, really cool.

Carolina Posma:

Yes. So it's also links already. So this is, uh, yeah, already trained automatically. And this is just your base version. So this is the draft version of your AI agent. From here on, um, this helps you to easily at least, um, yeah, get started because that's something you already asked me as well. How do I get started or how does, does someone get started with, uh, with creating a prompts? Well, that's one of the more difficult things to, to, to start from completely scratch. Um, so that's why they, uh, they built this to, to make it easy to go from zero to one. And from here on you can, can make any changes. Uh, just answer your question about the website and that you can copy the code and then you can have platform, uh, instructions over here. So let's take a look at how you can change it. So let's first take a look at the system prompts, um, and how they defined it. So over here, they have to role, um, like I mentioned as well, lead generation assistants, uh, core responsibilities. Uh, this is the same as the goal. Um, what you see over here is that there are always hashtags in front of, of the titles. And if you're not familiar with that, that's, uh, something called markdown format. And markdown format is a way to tell to computers how something is formatted. And, um, in short, one hashtag means heading one, two, hashtags heading two. And this way the AI agents or the AI will be, uh, better able to understand the structure of your system front. Because if you don't have this, it'll be just one big, um, yeah, wall of text basically for the AI agents. And it doesn't really know how it's structured and, and yeah, how, how it's, uh, important. It should, uh, different parts of the, uh, product part. So, uh, it has core responsibilities, um, the tone of voice. Um, this tone of voice is based on your, uh, website. So the AI agents. Team that took a look at your website, and based on that, uh, it's decided on the tone of voice. So if you have a very playful website, then the tone of voice will be different as well. Uh, conversation flow. So it's, um, tells us, this is a little bit like the SOP that I mentioned. So what should the, uh, way or what way should the conversation flow? You can make any changes here. So if you don't, don't like it, then you, if you don't want it to ask for an email address for example, then you can make those changes, key guidelines. And here are some FAQ knowledge that it created already. But you can also change that over here in the knowledge base if you go to

Isar Meitis:

FAQ. Yeah. Very cool. So basically, like you're saying, going from zero to one is very, very easy because it does, it basically creates a vanilla agent for the main topics that they have, like customer service or engagement or content creation with all the instructions already built in, which also saves you the hassle. Of my question before, of what categories do I need to have there? So these are already there and you can go and just add and change and massage and now you understand, but the general structure is already built in for you and not starting with a blank sheet of paper, which I think is fantastic.

Carolina Posma:

Yes. Yeah, exactly. Yes. So it makes it so much more easy because, oh, I, I must admit, I, I, I'm getting so lazy because of this, this, uh, yeah, automatically generated from before this wasn't there and that you needed to start from scratch and really needed to write everything and test and change. But this is already making life so much, so much easier. You are even more, uh, lazy or want to try more. You can also click over here and you can start recording and just say what you want the AI agent to do, and you can have it completely unstructured. You can just say anything that you want and then also say, oh, no, actually I don't want that. And then it'll also, uh, create a prompt just like this. Uh, so in inside the prompts, so. Let's cont, uh, continue by adding it now to Instagram. Because of the time, I won't make any changes to the prompt, but if we're, uh, if you would really create an Instagram, um, AI agent, I would, for example, um, ask it to not collect full name and email because we're already talking on Instagram. Um, but um, yeah, make those change. Play around, uh, click always on, save and test them. Play around with how the agent responds. But let's edit to Instagram as an example. You can click on configure so you're not

Isar Meitis:

watching, there's just a channels section and you click on that and Instagram is right there, and you click on configure. And that's, I assume, more or less it.

Carolina Posma:

Yeah, that's it. Click connection. Then. Make sure this is an, uh, important aspect that the Instagram profile that you want to connect it to, um, is the profile that you are already logged into on your, uh, accounts, uh, on your browser right now. So this is demo AI agent. I also gave the Q profile picture. And, uh, let me, now you click on allow and then the connection is now being set up. This may take a few moments because it needs to make a connection with Instagram. So let's wait for a little bit, um, from the moment on that you are connecting it. Be aware that your AI agent will start responding from now on your D app. So that's something to be aware of. Um, in the chat manager of Flo, you can always look conversations and look, means that the agent won't respond to that conversation anymore. So that's, um, something that you, uh, should be aware of. Um, you can have different settings as well, uh, such as that. Uh, if you send messages inside of the Instagram app that the agent is aware that you as a human sends those messages, um, send, uh, you can have it sent two updates. So for example, let's say that you're an e-commerce store, you connected with Shopify, and you, um, want the agent to update saying, Hey, I'm right now looking into our systems to see if, uh, uh, if it's happen, uh, or. If something is happening, if I can do something, then you can have that updated here as well. I usually have this turned off. Um, load al messages for conversations. So the moment, this is also one of the data question, uh, answer to one of the data questions. So if you connect your Instagram, um, the floating software won't see, won't see all your old conversations, and only if you, uh, allow it. Then the moment that someone starts talking to you from the moment onwards that you connected it, it can see the last 20 messages already. So then at least the conversation will keep on going. Um, something that I like to use for influenzas is a random response delay because of course, an AI agent will start responding immediately to you, uh, normally, but that's a little bit inhuman. And I also help, uh, influencers who, uh, want to automate their Instagram dms but don't want. Their followers to know that they are actually automating it. Um, then you have a random response delay. So it will take any time between I think two or uh, uh, one or 10 minutes to respond. And you can also set working hours, for example, if you want your customer support team to answer during the day. But once AI agent to answer during outside of office hours, then you can set that over here as well. So these are all functions that make it very useful, um, to actually, uh, connect it with Instagram. Now let's see if our AI agent is going to respond on Instagram. Uh, then I need to quickly search for the, what's the AI agent name? Demo, AI agent. If you want, um, anyone, you can just start chatting with my AI agent. Demo AI agent as well. Uh, let me see. It's over here. Follow, and I'll start sending it a message just to make sure. I think those of

Isar Meitis:

you're not, or not watching and just listening, we just open Instagram with a different account so we can chat with that account. Uh, and then we could just gonna send it a message and we'll see how it responds. Uh, yes. And we turned off the random delay, so it actually answers, so it actually answers immediately.

Carolina Posma:

I hope. So. Let's see. It'll take, uh, it'll always take a couple of seconds. Why? Um, because in Instagram it's very common that, um, people are sending multiple messages or sending. One big message for them. Send it in multiple messages, right? Oh, yes. So this is a, a built-in delay to make sure that, let's see, well, to multiply, uh, sorry. To make sure that, um, all those messages are combined sent to the AI agent, because otherwise it'll respond to every message individually. And that yes, gets a really weird conversation. So it's working. Welcome to Multiply. We're excited to help you to spur your business with ai. What brings you here today? And now we can also ask, uh, what's happening? Oh, I think that's okay. I think Have voice

Isar Meitis:

pipe somewhere.

Carolina Posma:

No, what what happens right now is, uh, the AI that I'm logged into right now, and I need to stop it, uh, is, or the Instagram channel that I'm, uh oh, has its own AI channel. It also has an

Isar Meitis:

so there, so the agents are talking to one another. This is the end of the world that we know it. So let's, let's pause here. I think, I think this was great from a demo perspective, I want to do a quick summary. Uh, and, and kind of explain everything we talked about and then I'll let you tell people how they can find you and work with you. But to, to summarize everything we talked about, one is we talked a little bit about what our agents and how they're different. But the most important thing is if we ignore all the explanation part, the process to create an agent and put it on, in this particular case, either your website with a code that will be copy paste, or on a channel like Instagram or Slack took maybe five minutes, right? Because we give it a link to the website and then we waited and there was a first draft of the agent ready. All we told it is what is the general role in this particular case, like a customer service one. So it built a system prompt for you. So if you know nothing and you want, okay, I wanna start at least experimenting with this five minutes later. You can have an AI chat that knows how to work with different tools. In this particular case, we didn't connect any, but we could have, uh, so let's say I have like a way to book stuff on my website or, and I want to connect it to the back end of that or a CRM or anything like that. Uh, literally five minutes later you have a good starting point and it's extremely powerful if you want to get started quickly and without any previous knowledge. And the other thing is it creates a very solid start for the system problem. So you can, from there, just you kind of understand what's going on and you can just add your stuff. And you don't have to be an expert on creating agents. Uh, Carlita, this was fantastic, like really a very, very well thought after a introduction and then the actual demo, if people wanna follow you, learn from you, work with you, uh, connect with you, what are the best ways to do that?

Carolina Posma:

Yes. So, um, you can find me on LinkedIn, Karina. Um, also on YouTube, I create, uh, videos on YouTube also making easy to follow explanations on how to get started with, uh, AI agents and my event to, to share that. Um, this Monday I'm launching the AI Agents Made Easy School community. And, uh, yes, I'm very excited for that because the goal of this community is really to, to welcome anyone who is, uh, doesn't consider themselves as, as technical to, uh, come join us and learn about AI agents, learn about ai, uh, automations, and really go from beginner to winner with AI agents and automation. So everyone is welcome, even if you feel like you have no technical experience at all.

Isar Meitis:

Fantastic. Uh, thank you so much. Thanks everybody who joined us live. We had multiple people in a very active audience. Sadly for you, you missed that. But a lot of good questions and conversation on, on LinkedIn, on Zoom you were at least able to see. Uh, but thank you everybody for joining us live. Uh, just for, for your knowledge, Carolina, we had people from, uh, uh, let me see, they said in the beginning, but from many places around the world from, uh, Singapore to India to UAE to San Diego, Cape Town. So you had a co literally people from all around the world, uh, in your session. So this was absolutely fantastic. So thank you everybody, uh, for joining us, uh, for being active, for asking questions, for chatting with, uh, with me, and with each other in the chat. And thank you so, so much Carolina. Again, this was absolutely fantastic.

Carolina Posma:

Thank you so much for having me.

People on this episode

Head Shot

Isar Meitis

Host