Unknown

hey there everyone how you doing another day in paradise I'm looking at the chat here Romania of Brazil Pakistan urugai damn we have people from all over the place la I love it I love it so cool by the way can everyone hear me all right just making sure that we do a sound check um video check everything is good if it's all right letting me know in the chats I can see our comments popping up uh I'm new to this thing so um I think this is the first live that I have done in

kind of like a while in kind of like forever so if anything this this be helpful all right Toronto Canada oh so good I love this I love this we have so much fun stuff to over today uh I was I was very excited about this and I thought about doing some crazy stuff I thought about maybe us doing like something from um something from my studio I thought about getting more of the team in the camera but in the end of the day I was like you know what this is the first day let's

keep it easy let's keep it simple and uh and just make sure that we keep going is as you folks have questions make sure to send them in the chat uh we have a few people that are moning the chat for me so they're going to basically pick it up what are the best questions so that I can answer later this is going to be very very very excited uh let's let's get a few things going I'm going to I have a few presentations to show and I want I want this to be more of

a chat so I want to basically walk you through AI agent how I'm thinking about it how the Market's moving about I want to talk about use cases I want to show behind the scenes a little bit on how we are using that crew AI I feel like there's a lot of things like that uh a lot of like conferences that I go to and and talks that I give that I mention that hey we are using this this is super exciting but I I want to show you like how exactly like we are doing

it how exactly we're using this and I have something that I'm so excited so excited to tell you and that is that this week we're going to launch a bunch of stuff so for everyone that has been asking this week there's going to be a UI and I'm going to show you I'm going to show you how you can use it we're going to use it together so that's going to be super exciting uh we are going to show you uh yeah this is recording I'm seeing some of the messages in there this is recording

I think it's going to be automatically in YouTube uh I hope so cross fingers if not we're going to figure it out all right so I'm going to show you a UI that is going to be great I'm going to show the platform we are going to launch the platform some point this week uh so we stick around for you to know uh but that's that's very exciting uh we're doing some final testings make sure that everything checks out uh and there is and there's also so much that I want to tell you about like

things that are going on if you saw the news today you know that c now is partnering with IBM I'm so excited I'm so excited about all this all right enough enough enough is enough let's get things going let me move here something to my other screen so that we can talk about it uh oh and also another thing is I got I got a couple cover set up so we can do this as well if it works I guess it works it doesn't works let me try boom all right let's go back to other

camera there you go all right so I'm going to share my screen first hello great news IBM partnership yes I'm very excited yeah IBM is great honestly I think they're cooking so much stuff and if you're thinking about AI IBM is kind of like OG right like they Watson has been around for a while all right enough's enough entire screen boom let's talk about AI agents so the goal here is to tell you a little bit about what AI agents are how they how they work how you can build them how you can deploy them

I understand that there's people in here in different levels of their Journey there are people in here that are already more familiar with AI agents there are people here they are not uh so I want to make sure that we catch everyone up real quick but at the same time we also don't like we also don't spend too much time kind of like covering things that most of people know so I'm going to ask you to be patient so I I catch you everyone up and then from that point out then we can then we

can dive into um into more more interesting stuff all right sharing my screen let me see if this will work [Music] now uh Ure that you have all the permissions that is fun share screen let me try this now and I have one final [Music] option there we [Music] go oh I see some messages popping up I love this I love this all right all right all right all right wellow numbers are numbers are pumping I'm seeing more and more people joining yeah I see people inant like yeah I'm trying to share the screen I'm

struggling with it a little bit I see some of you asking about it uh bear with me for a second uh uh uh uh it will be recorded it will be in YouTube later so anyone that missed it out are going to be able to are going to be able to see it eventually um I don't know if it's popping up when you click yeah H there know there's some message going up saying that uh make sure that you have permissions for sharing or something let me try this one instead we can do here all

right there we go let's do this so it seems like I can share a tab and you know what a tab is enough we we can do a lot with a tab there we go so again I'm going to start by catching everyone real quick so bear with me but I assume there might be people in here that don't know what AI agents are so just a couple minutes for me to catch you up so AI agents you probably have used llms so you're familiar with open AI you're familiar with and Tropic I mean unless

you kind of have been kind of like hiding behind a rock underneath a rock you probably have used llms llms are great and a lot of people use them but the thing is they're good to create content you can ask them hey create this email for me or uh create some something else like for me like a recipe or whatever it might be but then that's kind of like where just stop like they only work while you're actually talking with them and you're chatting with them but the cool thing about AI agents is that you

basically turn this and this is very simplistic you turn this into a loop so this llm is now talking with itself right so it's just basically chatting with itself or it could also be chatting with another llm so you can have another llm in here and now you have these two chatting with one another as well as long as they chatting with themselves and then you might be wondering like well well what what are they chatting about well that's the good thing the whole idea of an agent is that you can actually provide your llms

with a task right and again I'm oversimplifying here just to catch everyone up so you say like hey I have a task for you and you say I need you to do this task for me but then in order for this llm to be able to do something for you you need to provide it with tools so tools skills depending on the framework people are going to call this different things but let's say that there are tools for now so you say like all right there we go I want this llm to have access to

these twos and these twos they can be they can be anything honestly so they can be internal tools and a good examples of internal tools here could be um for example a connection with a database that maybe you have in your company it could be a connection with a CRM it could be a connection with an Erp whatever it might be or they can be external tools as well and this is kind of like more of the classical where you have like a a Google search or scraping or something else but but you have that

so the cool thing is this llms are going to decide given the task that you give to it when they're going to use those tools or not and when they decide to use those tools they're basically going to get this information back and they're going to use these tools back and forth up to the point that they feel good to give you a final answer so boom so you get a final answer here all right so this is basically agents in a nutshell very simplistic right when you look at them at this you're like well

this fil like very simple the problem is once that you bring this things into production things get quite complex right because you start to realize that you're like oh you know what there like every request that is happening in here I'd like to cat it because what happens if my agents are calling the same to over and over like I don't want to run out of like requests or heat rate limits or whatever and then you're like well I also don't want of my agents to hallucinate and to give me like a random m answer

and that becomes a problem as well and then you're like well I also want to give some limitations like I wanted these agents to only do maybe two or three or five requests at a time before they actually give me an answer or you can do a Time limitation so what you quickly realize is that a bunch of the things pop up so uh again just to name a few cash guard rails uh yes all right right so you got a quash you got a guard raos you got a memory and then once that you

start to have mple agents now you have delegation you have asking questions and not only that but you want this multiple agents now to be able to share that so you want the memory to be shared you want the caching to be shared so now you need to to B basically think about all this so where I'm getting is things can get quite complex once that you get to this stage and we're not even talking about how you orchestrate the agents especially if they're doing that in parallel or organic the one not organic parallel or

hierarchically but the one thing that I think we we got uh in crew that I think is so good is that we made it that simple for you to build the things and honestly this is the thing that I like the most about crew uh I got very inspired by like other Frameworks that I worked in the past things like crew rails and all that and some of their ideas of conventions over configuration in order to bring some of that into CI and that has proven to be very helpful all right so you folks understand

what agents are technically the reason why this matters is because it unblocks automations right forget the name AI agents for a second like don't think about it if you don't like the name AI agents you're find everything like it's too much hype or something just don't think about AI agents think about AI powered automations in these are alterations that were not possible before the reason why this is different and this is exciting is because these llms they come with a few things one because they can basically understand what is happening in retry things they have

selfhealing capabilities so they can self-heal if they do something wrong they retry it if they try something away and doesn't work they try again another way so this can be very helpful this ability of selfhealing as they go through the things the other thing is reason so they don't follow a straight line they don't like to do a then B then C there are times where you might want to do that but the beauty of this agents is that they can dynamically react to real time data and this is important because in a lot of

the internal automations especially on Enterprises and companies you're thinking about hey like if giv like a customer data I want this agents to behave in a certain way or to do something else or giving like another customer data you want to do something else and entirely so this can have a lot of impact around how your agents work and the fact that they can reason the fact of that they have agency the fact they can self heal the fact that they can basically have multimodels um they have a lot uh I'm seeing the chat here

is self healing not a little too much fancy wording I don't think so so the selfhealing uh for me comes from Elixir Elixir was like U has this kind of idea of self-healing and why I think this is important is because in the context of automations if you think about how you automated things in the past especially in in companies it was using something called RPA uh robotic process Automation in that when when people like got to know about that people got very excited but it fell short under the promises that it made because it

was very brittle so if a buddham changed a place it didn't work anymore if a process Chang a little bit or slightly then the automation didn't work anymore and with agents you don't see that because if they if they basically hit a dead end they can retry something um all right so we talked about AI agents I think everyone caught up on what AI agents are by now so why don't we move on to something else let me go over a few things and I think this is important um and here is me giving more

context to you all on the business sides of things because I think as as Engineers or as um Executives in companies doesn't matter where you are like if you're spending time or spending your company resources into something that is is basically taking resources from something else you want to make sure that this is worth it so let me go ahead and share my screen again uh real quick [Music] all right I just need to approve my screen sharing real quick all right so the thing that I want to talk about in here is the the

genie is not getting back into the bottle right I'm talking I'm talking about those like in some of the conferences that I'm speaking to and now you're actually seeing this out there so it's all over the media honestly like you know about this right you you're seeing like uh uh Salesforce for example doing a huge conference talking all about artificial intelligence AI agents in Salesforce is one of the biggest companies in the world so you know like if they're if they're putting money into this they believe it especially like with the whole agent Forest thing

that they have going on uh now you also have HubSpot the other big compactor talking about agent AI in explaining how agents are going to be important for marketing how it's going to be important for sales how it's important for the lighting customers so a lot going on in there as well and and then you go into uh and then you you go into like some of the news that you're seeing here like CL AI assistants Goldman Sachs looking into it AWS X looking into it feels like there's there's a lot going on so when

we started crew as a company initially our hypothesis was well 2024 is going to be kind like a a small Market it's going to be kind of like the early days right like you're going to have Pilots you're going to have like like early adopters people experimenting with things uh and I got to say oh boy I was wrong because things moved way faster than I expected and we have people deploying things in production already and I'm Amazed by that then we thought all right maybe 2025 all the way to uh 2030 uh there you

go presentation mode I think that helps well I I lose the chat when I do that so let me try to change something there you go this might make it a little better all right and then uh we are like well maybe on 2025 that's kind of like when things are going to start to getting into production and all that and and now like these are moving way faster um way way faster than I think most people expected and it's funny because I keep hearing people comparing AI with uh with the internet and I think

it's a good comparison especially because I think both are going to be kind of like a a good platform and what I mean by platform is like think about the internet and everything that you do happens in the internet right you watch movies on the internet we're connecting now over the internet you meet other people over the internet you order food over the Internet so internet became this platform where things happens or happens happen true and with AI it's a similar thing right but AI is going to be a platform so they going to meet

people they're going to be recommended to AI you're going to uh order food that is recommended to you by Ai and you're already watching movies they're recommended to you by AI so I think that's an interesting comparison that sad if you look at the doom boom and everyone that was talking about hey this is going to be big this going to be big there was a difference if you actually put a website online on the early days of the internet you had nothing to gain from it but if you actually do now if you actually

bring um if you actually use AI you can see results in like the next quarter depending on like how big your company is or how small your company is you can actually see results in like a week a month it's it's insane some of the results that we're seeing out there we have like we have some early customers that they are telling me this is not me they are telling me that they are saving hundreds of thousands of dollars by running AI agents it's it's a little insane and then if you keep going I think

in the future this going to be like and this is like maybe I drink too much of my own Kool-Aid but I think this is kind of like where things got interest right by by 2031 God knows what's going to happen but it's the trajectory is definitely going into the direction that you're going to have this entire ecosystems their entire economies and agents are going to be interacting to the internet and what they're going to look like and again this is not something that it's so far in the future because you are seeing the Echoes

of fit happen right now like crei just partner with IBM one of the biggest companies in the world to help take AI agents to as many Enterprises as we can worldwide and we're already seeing people using like crew in so many different aspects and now IBM is going to be offering services on that and help people adopting it and help people deploy these agents in the production environments even if they don't have the right skill sets and this feels great we so excited about this partnership uh I think this going to I think it's going

to be great honestly honestly and and then there's a few other things as I was as I was telling let me see if I can change the screens more rapidly here I want to show you a few other things so I mentioned to you I mentioned to you uh Salesforce there you go Salesforce one of the biggest company in the world talking about AI agents and then it's not only Salesforce I mentioned to you as well ch this out boom HubSpot the other big competitor also talking about AI agents so much stuff going on and

it doesn't stop there if you if you look it up and this is what I'm telling you about results and how you get a results so fast with AI nowadays is check this this was from the last earning calls from almart so I don't know if how how many of you are kind of like um uh how many of you are like into kind of like Finance or anything but less um less update from almart less earning calls one they blow off like they they blew up their estimate the estimates that they had on how

much money they would be making and a lot of it seems to be due to AI they're using Ai and automation to drive efficience and they think that they can get things to be aund times more productive by doing this so this is what I'm telling you like this is the difference right with with the internet early days you basically getting a website wouldn't necessarily impact your business from the day one but with AI you're seeing like this crazy results crazy results that people are getting and there there's yeah analysts saying that they could probably

save billions in operational efficien in the next years so this is huge again forget to name AI agents for a second we're talking about Ai and Automation and efficience and uh and and that is just that is just so much all right let me go back into boom boom uh there's a few things though that I believe it for that becomes a reality there's a few things that need to happen right and this is where I think people are not realizing it's not it's not simple right oh this is so good it's so good there's

so much potential yes there is but it's not a simple thing A lot of people are trying to build like different versions and kind of getting burned um SS are doing like in a similar scenarios and that are working but I believe there are at least four things that needs to happen for AI agents to go mainstream especially in organization uh and I'm seeing the chat here an AI agent system that can automatically create train and grow a organization then yes I think so I think I think you can get I think you can get

AI agents that are self-improving you actually have some people working working on that uh we have a version of that in our platform I can actually show you um well let's take the platform more towards the end I want I want to talk about some things so four things that I need to happen for this F this future of AI agents to become a reality in my opinion one needs to be production ready so whatever platform you're using whatever tool you're using or even if you're buing something yourself internally whatever you do needs to be

production ready and what I mean by that is not like oh it needs to have tests no no no it needs to it needs to run on mpo clouds it needs to offer a level of control a very fine grain control so you can say who get access to one it needs to have a way for one person to kind of like register what are the approved vendors what is their proven llm provider what is their proven tools to be using so out that it's not even AI features like those are like traditional Enterprise features

that allow Enterprises to actually adopt technology so if if those things are not in there there's no way that companies can actually adopt this um then I think also whatever platform that helps companies basically fulfill this Vision needs to be also Universal and Universal here here is people don't want to get vendor locked right we all saw for example a few months ago or maybe a year or two ago when uh a big Fortune 500 company basically spent millions of dollars fine tune GPT 3.5 just so that GPT 4 would be launched and would be

better than their fine tune without any training without any fine tuning so it's kind of like it's kind of like insane to see that the because the attack is moving so fast you actually don't want to be lock it to anyone because you don't know what's going to happen so you might want be using like open AI U llms today but then tomorrow entr Tropic launches a new thing you want to be using in Tropic and IBM now is launching a bunch of like open source models with their gradient models pretty good so maybe you

can change to that so the other thing is needs to be Universal it needs to be easy for you to change from one model to another model from one provider to another provider from one integration to another integration so again two things so far production ready find grein control Enterprise ready features Universal connecting to all those llms connecting to all those Integrations the third thing is it needs to support complexity so llms by themselves cannot automate processes and traditional graph-based processes they don't deliver the value right what do you want to have is you want

to have something that can self heal that can react to this data live and I'm going to show some examples real quick uh showing you like on how this can on how this can be helpful um and Paulo here send a message in the chat saying agents are as powerful as the llms that they use yes you're correct depending on the models you're going to get an agent that is more capable of reasoning or not but remember that powerful here can mean different things you don't necessarily want a big LM sometimes you want to optimize

for Speed right so you can think about hey how speed compares to kind of like the quality so there's a few different things in there uh I see also uh CD Badu mentioned HubSpot partnership school I have seen darash yes so uh darash and I are actually close friends uh I mean I like to believe so maybe not so close but we are friends uh we chat over Mayo quite a lot and he's actually one of the investors in crei so if you didn't know that yes uh their mesh has invested in us and I

love to have him as an investor he's such a nice guy all right and then I save the best for last it needs to be fast and easy this like out of everything that I mentioned so far this is the most important thing and I'm going to tell you why this is the most important thing because if takes too long or it's too hard for you to build AI agents that is screws UPS the economics that is screws UPS the economics because now you got to deploy I don't know two three five 10 engineers and

a product and a design person to build this for three or six months and that limits your ability to innovate and to move fast and and remember the Market's moving fast you don't want to be stuck in time building whatever you're building you want to make sure that you're moving fast as well so what ever platform that allows these things to happen that allows companies and Enterprises to adopt AI in AI agents needs to be fast and easy because people can throw things in the wall see what it sticks and then double down that right

and honestly crew AI checks all of those boxes especially on the product and on the platform that we are going to launch later this week like it checks all those boxes and it's it's pretty good because I think this one this one for me is the most important piece because think with me if you make it so easy and fast that people can de like even people that are not Engineers can deploy and they can deploy super fast then they can just try whatever and then what works is you when you double down and that

puts you in the position of strength where you're actually playing through the market movement you're playing to the market kind of like Visual and like what is happening so uh so yeah it's it's pretty good I see a question from Jay Louis Jr will there be a visual version for the corporate use yes there will there will be I can give you a nickpic today I thought about saving the platform for kind of like Wednesday for us to use together um but I I can show you as Nick peek and we can go over it

and maybe we use it a little bit today uh and maybe a little bit tomorrow and then Wednesday we go like full deep in but yes I think like you got to move fast and honestly I think this is also I think this is also about the companies right so crei I think has an unfair Advantage because we move fast and we need to keep moving fast because the Market's moving even faster so all right these are the four checks then if you look at what we have today these are kind of like the the

levels of automations that we have right you have kind of like traditional R base Automation and this is more traditional RPA for example um then you have kind of something that is a little better uh where you have kind of like some rules mixed with Gen and this is where you have like graph based automations or node based automations kind of like that if this then that kind of like Automation and you can have some gen here and there but it still like it's very rigid right it's very rigid you you may have like some

some control here and there but that that kind of like it's it doesn't have the agency so it's not an agent um and then if you look at here you say like all right you have complex llms now with open ai1 preview and I did a few tasks with it I got to say it's pretty good and already works with CI so you can already use ow and previl even though I wouldn't recommend the results are amazing the quality is so good but it uses too many tokens that take too long so probably like a

couple away from being something that you can you can deploy for gent accuse cases then on level four you have kind of like where I believe create IES right now where you have um MTI agent and where you have dis ability of flows where you can use the flows to connect multiagents together and connect one thing to the other so that that is very exciting uh and I think and where we want to go at least with crew AI is I want to go to level five where it's fully autonomous right where now you can

have your agents basically you don't need to give specific tasks you don't even you don't even need to Define your agents basically they alone can understand and figure out themselves what are the agents that they should be what are the tasks that should be and kind of like self organize I think it's a little too early from that we are doing some experiments internally but I think like whenever we get it here things are going to get pretty pretty pretty insane uh I see a few questions asking uh is it possible to create an automation

where One agent handle WhatsApp Integrations another one process request via apis yes yes yes yes that's good I can definitely I can definitely show you all how you can do it uh all right so given out this uh we we have been building Crea you know that and we have been building a product version as well so the version that the product that I have been building is pretty good uh I'm going to run a demo uh and I'm going to show you some of the features eventually uh I won't spoil you too much because

again I want to show more in depth some of these features tomorrow but I want to give you an overview there's a studio feature so you can build with no code there is workflows that you can have in there there's ability to have humans in the midle there's ability to manage manage crew there's ability to T Cruise there's examples there's a lot going on in there so I want to make sure that I show you that you know I think that's going to be pretty good uh let me see what questions we have in there

um how do you see by comparing with auto Jam um honestly Auto gem Is AO gem is okay when when I build crew uh Auto gem was already around I was in Need for like an building agents and I went a framework and I saw autogen but it didn't work for me I I thought it was too complex uh was not I was looking for I was looking for something that I could deploy that F more production ready and I just didn't feel that that was it I really liked uh a few other projects that

was a chatav I love the idea of how chatav organized some of its components but then the problem is it was not meant for production right it had like a game like UI and I was like all right this is not what I want so I was like all right I'm going to I'm going to build my own thing and and that's kind of like how we got to start building crew um Ashley asks can you talk about crew architecture cognitive acori there planed for developer better agent yes for sure I can definitely talk about

memory uh right now uh right now agents and CI have three types of memory they basically have uh long-term memory short-term memory and entity memory uh one of those is basically like a a rag pipeline behind sces the other one is a regular like database and I think the third one is actually a regular database just a local database um and that's kind of like how you can like use that um and yeah I hear Kenneth I I I also had a hard time with auto Jam honestly I mean at first I was a little

like uh frustrated but what I learned later is that they came from like a research project and I was like all right this makes more sense like this GES like as a research was part of a paper so I understand better like why they take some of the decisions that they did uh and and that's fair game uh what are the key difference between the Enterprise version of c and the open source oh that's a great question let me see if I have a slide in here all right so when you talk about uh open

source we basically help you plan and you help you build right because you have so many features like so many content that we have been recording and then for building you have the framework as well that you can use to actually build those things with Enterprise we heavily expand the planning with with like templates more content and we heavily expand the build because now you also can um you can also sorry because now you can also do the no code the crew studio and all that you can deploy them automatically turn your agents into API

you can run them on schat you can sa rate limits you can do a bunch of that you can automatically export a UI you can monitor them and I'm going to show you some features around that where you can monitor the quality their execution there's a lot that is going on in there and then you can intergrate on them you can test you can compare different llms you can you can do a lot of in there all right enough enough enough enough let me show you the bad boy real quick uh I think this is

going to be interesting by the way bear with me what I'm going to show you is something that is running localhost um and I think you're going to I think you're going to like it let me answer that one question before uh is z Kiki why should users choose crei from Lang graph except from the reason of feing AI agents very fast well I would say link graph and CI they're a little different in a in a lot of ways like link graph feels very much like more if this then that kind of automation where

you kind of like you have the ability to do some of that on crei you can also do that with flows but I think the ability for you to actually have agents like uh L graph like what you have is like the all this like traditional coding logic and you can call llms but you don't have like the self healing of the agents you don't have all the features that come with the agents especially when you go MTI agentic and my understanding as well is with with lra feel more focus on developers a little bit

uh and we are more focused like at least on the on the Enterprise product we're more focused on Enterprises so there's a little bit of that uh but we can talk more about that later let me share my screen now another tab Boom by the way let's before I share this let's do a quick break let's do a quick break I want to I want to I want to go to another camera I have another camera in here it should work now boom did it work ah you see this what this is what I geted

when I try to do to be fancy and have two cameras never mind let's let's go straight to it gray eye anyone be with little orientation yes I agree I honestly I I think to go back to those four checkpoints I think CI is the simplest and the fastest out there I don't think there's any easier way for you to build AI agents now so there's there's that but let's go into the platform real quick and again this is local environment so uh bear with me hopefully it works first time there's a few things that

might be broken in my local environment has been a lot of building and not a lot of sleep but uh I think it's going to be very good all right so in here you're seeing some some of the features uh there's a lot of things that are uh heavily uh gated in here I can show you the production version in a second and you're going to see there's more things going on in there but the first thing that you're going to notice is the ability for you to deploy your agents either from GitHub or you

can and also do from basically a zip file if you're not using GitHub once that you deploy them you can see them in here and they become an API and this API is production ready what I mean by that is for the more technical people out there this is private VPC with SSL with auto scaling bare token like most of the things that you would win from here you could actually come in here and Export a react component it going to actually allow you to basically export an entire UI and host a UI in here

uh before I go into specifics let me show you what you want to see and that is crew studio so crew studio is our way for you to build agents with no code and honestly I love it I use it all the time so the way that it works is you can start by chatting we put a lot of thought into this where you start by chatting with crew in order to create your agents and then as you get more comfortable with what you're building you can basically graduate into more complex as UI and you

can see how this is way more approachable for people that are non Technical and way easier and faster for people that are non technical than just jumping straight into a more complex no code UI so in here for example let's say uh I don't know let's start with something simple uh um all right I'm going to do something that simple right like that we can get more complex ones let's do um create a crew that is able of creating uh blog posts for crew AI whenever we could a new version based on the release notes

I can share with the crew there you go so kind of like a simple use case um I have tried some like way more complex stuff and they work fine maybe they're they're in production I can go into production and show you some of that let me see if I can actually load production here while this uh gos boom all right let me go for production real quick so that I can show you that nice all right so this is the production there's you can see more features in here more things as well there's a

few different Crews that are deployed in here they're running there's one that's actually being deployed right now and also this UI is not 100 up to date but in here you can see some of the crews that I have been creating so I say like hey I'm going to be talking with some reporters I need you to help me research all their latest posts their content do a few analysis I basically to understand standard perspective on Tech ai ai agents anything that might be relevant for my interview then it kind of creates a plan for

me like gives me an output say these are inputs these are agents these are tasks I ask you to change a little bit it updates the the basically the plan and once that gets to this stage and you feel good about it you're like all right now I understand what I'm doing so you can actually click on generate crew plan and it's going to give you this table like UI in the stable iqi it's kind of like a step up from only chatting but you can still changing things you can come in here and like

all right this is not a senior there's going to be like something else in here so you can keep changing your agents now in a little bit more complex UI but nothing crazy because you already understand what is going on you know what the plan is and once this feels good you can click on create crew and then you go for a more node likee UI and this nod like UI you can actually get a more technical in here and you can start do things synchronously if you want to and you can basically update your

tasks you can update your agents you your tools you can look at all your agents in here you can update the models that they're using so there's a lot that can go in here and then you download the code if you want to or you deploy it right away and and this is pretty good because ability to deploy the code the download the code for me is is great because this means that I that I'm okay writing code I can basically use this to step up do super fast AC crew and then export it and

take on from the code so this helps me basically build so much faster than just like uh just like starting from the GGO so there's a lot going on in here and if you click on deploy then basically that goes into that becomes a crew in here oh by the way I clicked on the temp plates by mistake so there you go a bunch of templates uh all right let's see so This Crew is deploying that is going in here you can see some of the crews that have been executed uh and once that you

get in here you get your previous executions you got to train your agents live so you can actually enforcing training in here you can monitor executions you can test your agents and compare all the different models that you want you can get a metrix over time you can see your logs you can update settings you can set alerts on performance so if your agent starts to kind of like not perform that well or execution time there scheduling there's there's a lot going on in here and I would love to dive into some of this with

you all and show you how you can use it and I didn't even talk about the management UI and that is pretty sweet and I didn't didn't talk about some of the more Enterprise features like the LM connections environment variables or Integrations there is so much that you can do in here with like some of these features uh I'm going to show you like just one example like a simple example because today is the first day right we're just parting talking about agents we're going to talk about more agents tomorrow and Wednesday but before we

get to that I want to show you one thing that I I find pretty exciting let me get one thing set up so that you can you can now tag along hopefully hopefully it allows to share my uh my screen this time because I I would love to show you this this is so good all right let me try all right so let me try to do one thing here let me share screen and now let's hope this works I don't think this is going to work ah so unfortunate ah I I wanted to demo

you how you can actually well I think I actually can let me try one thing uh just bear with me for a second because I think you're going to love this one all right all right so let me show you this again boom all right so I'm going to show you one thing right here you can see the the the crews that are deployed one of the crews in here is one that I like to use quite a lot and that is the prep for a meeting that play So I nowadays I have a lot

of meetings and a lot of thems are kind of like last minute like last minute meetings I'm like oh I need to jump into a meeting uh I I don't know like I I don't I don't know context who this person is and thing so what I did I was like all right let me vi a crew that can actually research this person and give me like notes and kind of like prep myself for the meeting a very simple use case right uh some of the use cases that we're seeing out there like they're way

more advanced than this and tomorrow we're going to be talking about this use cases we're going to talk about how companies are using this in production and saving like hundreds of thousands of dollars now but hey check this out these are the executions in here I'm going to show show you how I can actually execute this directly from boom [Music] pom there we go all right so I'm going to share my screen again so check this out this one is cool so because PL create Enterprise has a connection with slack I I can actually kick

off crew directly from slank so I can come in here I can click like kick off for crew it Leist all my crw are in there I it's like all right the one that preps for a meeting and then it ask me for like some context I'm like all right like what what is this meeting about and and this is basically what I told my crew the context would be so I'm going to say all right I'm going to have initial uh sales call with zenas they want to use agents um this is their email

and I don't know if that's someone's emailing there or not these are the domain and there you go kick it off so what happens now is like these agents are going to start to do their work and we can we can even tag along that work in the platform but once they're done they're going to send a message to us in here and going to be able to see the res right in here and this is just one example uh but this is kind of like this is this is super cool and we are not

only support Zack we support way more stuff as as well so I know one of you actually mentioned HubSpot so let me show you this another way that we use crew internally as well is in HubSpot so you can actually have a workflow and then you can kick off a crew directly from HubSpot you can like select whatever crew you want so in this case I can hey I want to use the lead score in 10 plate and then you can actually do something with that so in here I'm saying sending like uh basically a

notification so that you get a message back on with the crew resumed um directly on slack so in here you can see another possible Integrations and and yes you can integrate with so much stuff actually on the on the on the Enterprise product we offer also an integration with zapier so through zapier you can actually connect it with anything so here check it out uh we got the answer in laack so this is the person that I'm meeting this is some information about them there's like the social media presence some background of zenas all right

recent news all right launch Adventure fund this feels good introduce new AI features including AI agents that feels great financial performance Revenue growth damn that's great all right so I can jump into this call and say like hey I heard you folks are doing great amazing report on the growth that your folks are seeing on the revenue I'm very impressed about that and I also heard that you launched AI agents for like workflow automation so this feels great and you can see how this helps me like just jump into a meeting that I don't have

context again more of a silly in like automation but I like to show like the potential that can come from some of this now tomorrow tomorrow we're going to talk about complex use cases tomorrow we're going to talk about what are the use cases that these big companies are actually using out there and we are working with like some crazy companies honestly there's one of the companies that we are working with that they basically have Crews that are getting live feed of live games and they're cutting editing out adding adding sound and captions and posting

that in social media live as the game is happening and they are using 3i for that and that is just like insane we're also seeing uh legal use cases so we're working with a big Telecom company there's basically using crw to kind of like take a lot of the load from legal and now legal only has to do a few approvals so they don't need to spend too much time as they used to in the past and then other than that we're seeing companies doing marketing uh we're working one of the biggest uh real estate

companies in the US where they basically are using crw to monitor legislation across the entire us you to create content so we can talk more about those too there's so much that we can do and I didn't even talk about coding use cases because agents can code as well and there's like some incredible content coding use cases out there so hey there's a lot going on this week is AI agents week we're going to talk about AI agents today we're going to talk about AI agents tomorrow and we're going to talk about AI agents Wednesday

I'm going to use the platform we're going to be some AI agents tomorrow we're going to get our hands dirty we're going to talk about complex use cases we're going to talk about some of the companies that are using this and then on Wednesday I want to you to be it so on Wednesday is all about helping you build this AI agent ourselves and you can use the CI platform if you want to we're going to figure something out bear with me it's going to be gray but this is going to be this is going

to be a very interesting so I really appreciate you all showing up my idea was to spend like just one hour for us to chat for us to like start talking about AI agents I hope you have liked if you did share this with people share this with people from within your company share this with people that you know share this with decision makers a lot of people are feeling the pressure of like oh everyone's talking about AI agents what are they how do I use them what are the potential use cases so if you

have anyone that you think should be in here invite them to join us next um next session tomorrow because then we can actually talk about the use cases we can actually talk about how to build them you can actually talk about Roi you can talk about all these other things so make sure to bring this people today was all about setting up the context talking about what are AI agents how they work how you can use them talking about how the word is already embracing them this is happening no matter what F companies are doing

all the biggest companies in the world are already doing it like this is not going back and then in order for that to become a reality what do you need you need to be Enterprise ready you need to be production ready you need to be easy and fast to use and I think honestly easy and fast to use is going to be is it's going to be like something that we are unbeatable on that uh and not only that also I think like are the most Enterprise ready and there's so much more in there um

about access I know I have seen some messages people asking about access to the platform I promise you that you're going to get access to the platform this week this is happening this is happening this is the week so today we announce the partnership with IBM Tomorrow there's another announcement I'm very excited about tomorrow's announcement but there's an even more impressive announcement on Wednesday so I would say stick around Wednesday is going to be insane it's going to be the biggest the biggest of the announcements but we're going to do another announcement tomorrow another one

on Wednesday thank you so much I really appreciate I really appreciate everyone that uses CI I really appreciate everyone that is showing up the lives I'd love to see us I don't know how many people join this I'd love to see more people join so invite your friends invite people from your work invite decision makers invite people that you know that are thinking about this I would love to also share some of the news with them and there's more news tomorrow all right thank you so much folks have a good night I catch you later

have a