Introduction

problems so incredibly sort of in the middle here uh we have this stability of algorithms but we also have the flexibility of the intelligence so i hope i explained this

clearly clear enough if you have questions guys by the way uh feel free to you know shoot them in the chat and we perhaps take them a little bit later so we don't forget about them um

Platform Capabilities and Integrations

um yeah this is essentially um more you know we have a lot of integrations as i said 500 plus integrations and you have hallucinations on data and we can also take you know actions is parallel

like a lot so if you look at just give the example and if you connect an app it can only one action per time we can have thousand ones in parallel so it's actually a huge difference in speed as well uh yeah and this is like the elegant probably problems i talked about a little bit

Function Calling and Benchmarks

um and we have actually done benchmarks on this we compared or um to be clear this is for function calling specifically uh what function calling means is that the ability for ai to take actions that's all it is that's what functional callings uh mean like how uh platforms like chat kbt can

use external tools or apis uh to help you with your problem and we are actually beating these other models uh in this specific area uh so because we have a genetic model it works very

very different as well like actually we only have one model right now but we have three different models that you know have different abilities um

Security, Compliance, and Trust

essentially we also have compliance we actually are we have the audit in january about ice compliance because we um yeah it's like uh very important one thing recognized during our time like building a platform is because you know people have a hard time trusting us as a company

or the apps that we connect so this like makes sure that people actually can trust it the world's best standard for data protection so we take we have non -tolerance when it comes to data security we have the highest standard in everything basically

but it can it can still be a big note if people don't want to correct their data like that's

Developer Experience and API Usage

understandable still it is just a before those code geeks out there it is like how you can use the api essentially it's extremely easy to use the incredible api and to perform um yeah stuff with

it um i could go back to the model as well to the platform as well here and

Live Demo: Data Retrieval and Visualization

we can actually do one more thing that i think is pretty cool um we can actually let's do the volvo use case well we can take another car company let's say um polestar get the financial financial and sales data last through three years maybe let's see if it's fine it doesn't always

find it but that's fine you know if we as humans can find it the ai can't usually find it either but sometimes it finds of course things that we humans have very hard time finding but again it's using the web search to perform this here we go so i got a bunch of financial

data here and now let's show some charts so you actually have built -in charts in this bathroom as well so it can look pretty neat if you don't want to go to another platform to view the search

a chart as well so now it's like looking for tools to create this and here we go so here i got a bar chart and yeah over here line chart

results yeah looking good and i can download these ones etc looks pretty neat however one

Social Media and CRM Automation Use Cases

of the biggest use cases we have i would say it's linkedin so you can actually post on linkedin automatically with this to perform stuff um i can show the app i don't need to do it but

we also have apps like uh i think instagram is yeah instagram you can for example make a work um a use case where you out to answer uh people in a chat so if you are a company i get a lot of you know requests you can make a you know both answers then instantly and you can also connect

like databases and we have for example super base if you want to do financial analysis like analysis on your data or something like that you can use super base and you see all the like we have a ton of different you know categories of um uh of uh professions i got the question here

Recruiting on LinkedIn: From CVs to Candidate Outreach

let's talk about linkedin please for finding your roles originally that's a great question because actually uh we have a lot of customers specifically for linkedin and recruiters primarily

Primarily, recruiters, we have like two use case for primarily, it's partially finding people at different companies to reach out to, like candidates for companies and also reading their CVs, right?

So we actually have a flow where essentially they get like a ton of CVs right into the inbox at this recruiting company so imagine you're getting a hundred cvs and you need to

read through them in all these different formats first of all we format all the cvs pdfs usually to one single format so it makes it easy to see it but also we figure out who is the person on on linkedin so it actually i can actually show an example

so if we have i don't know if you have a cv actually i can take this invoice for example i think it's fine um so i take this invoice and we actually have a file support as well so you just upload files and it can read it so in this case this customer uploaded a bunch of cvs and

and then ask it to, we can actually have LinkedIn as well. Yeah, I need to connect it as well. And as you can see here, I haven't connected before. So now it prompts me to connect to LinkedIn with my account.

Hopefully I can provide my password here. I'll log in and I allow it to access my account. And here we go. Continue, add to conversation, and here we go.

so now i have linkedin in my conversation which is great and now i could this recruiting company again they asked read all these cvs and then find their linkedin now of course uh this wasn't the cv it was an invoice they didn't want to find it but you get the idea

Document Processing and Order Automation

this would have to work we could actually just try to read this invoice so another use case is that we it isn't a customer yet we actually talked with a couple of them with this it's about it's

about automatic ordering like essentially when a industrial company get a bunch of orders they want to you know fetch all the pdfs from a gmail and in the background and then we read all the pdfs and then we put it in into their crm system and their order management system so it gets in automatically otherwise right now people manually read each pdf and manually input the data in the

crm system which is yeah as you all know extremely labor intensive as this is like a um yeah and it actually tells me how to bookkeep this as well which has been nice um we have actually another

Bookkeeping and Operational Efficiency

use case when it comes to bookkeeping essentially there was this company that had a lot of problem and with specifically bookkeeping they actually used our api in this case or you know our tech behind this platform to create their own bookkeeping system so basically they cut like 30 percent of their you know repetitive tasks which is pretty amazing when you think about it

so yeah that's uh yeah i i can't think a lot of ton of use cases but you know um

Q&A: Job Boards, Privacy, and Getting Started

now i forgot to read so so you can search a job board um yeah it can 100 percent francesca so you can you know go to this like uh groups and find posts and read it um um read read through all the you know threads and find relevant people um francesca also write wrote um if it

it like gives you ids or if it gives it prompts um so if i understand it correctly uh uh it's a if if like incredible give you uh um yeah your description okay got it got it uh so job description yeah 100 so they give you the job description of the linkedin um i i don't want

to show the linkedin because privacy uh reasons uh i might pull people into yeah we take privacy very very seriously so that's why i don't want to share public here but uh that's all right because you guys can go in yourself into incredible and just try it so if you go to just open up a new

window and you go to incredible .one this is the website so essentially uh you can you know view all the apps etc here uh but you need to sign in so you sign in with google anything add your linkedin connect your account and you're on you can just you know uh do whatever you want

just play around with it and that's the beautiful thing here right uh just play around with it and try your you know your own workflows and your own uh prompts and that's the like the funnest thing to do and just you know from people i have had meetings with uh there are so many

different use cases you can't imagine with this platform so we have made this pretty you know generalistic platform which basically anyone can use it's super simple to use and hopefully you know everyone will um find their specific use case solved it's awesome uh do we have any more

questions by the way i know francesca is right now going to try the the linkedin uh which is awesome well what do you think you uh what's the time um we've got a little bit over time but that's okay like i said we've had dropouts and it's all really really good value um i was really impressed when

i came across this um i think i'll answer francesca's second question um and then maybe i'll take a i'll give you a question then close it there yeah and the second question is um does

Generating Ideas, Prompts, and Job Descriptions

it give you ideas or prompts like this is my jd uh this is my job board i guess how can you make my life easier job description uh yeah yeah 100 so it can like uh it can take your job description and find jobs for you uh if you're searching for jobs but it can also help as i said before

recruiters if they're searching for certain people they will take these job descriptions and find them and the beautiful the beautiful thing here is that it um if you're looking for a person a company look for a person uh their job description might vary uh like if you're uh you're searching

for an engineer that can draw you know cad and draw um like mechanics in the computer and like there can be a hundred different synonyms to this description so the ai will actually you know come up with these different keywords and find the right people automatically through the linkedin

Q&A: Handling Hallucinations and Model Design

so yeah that's awesome and then my question is about hallucinations can i just ask you to re -explain how it is that um you basically don't have to deal with hallucinations as a problem

yeah yeah sure now that's one of the i love that you asked that because that's like one of our most fascinating uh techniques out there but essentially again like these traditional models they think in English they out tokens they hallucinate out they hallucinate out

English language right which can be an accurate really we have created models that live think in code so it writes code and it access the data you give it so it means that all the data you see in this platform it it's accessed uh purely from the source itself and so it live thinks in algorithms to access the data in order to you know accurately provide it to you yeah so that's like yeah i'm

gonna i need a follow -up question because this is fascinating and so so important um you said don't think in tokens but do you think in code is is that is that strictly right or is it using next next token prediction just within a coding language yeah that's 100 I actually love that uh that the saying of it that tagline uh I might actually use myself like we don't think in tokens

and we're thinking you know uh in code really uh so that's 100 true yeah so it's if it's not using next token prediction you know how how does it work as much as you're allowed to say I I'm not

engineer behind it so i know i don't know like the specific specific details i have to be honest uh it's like we have a like a bunch of huge team of like researchers and developers creating these models and i myself or a developer to be clear like i developed since i was like 15

years old but you know as a founder you take the role or responsibility of other things as well like product management and all this kind of stuff so usually now i do sales and these kind of presentations etc and talking to customers i love talking to customers by the way i love the

reactions of the platform overall um but uh yeah i don't know the specifics i really don't yeah um okay uh well it's it's just obviously such an impressive advertisement that you don't hallucinate

Wrap-Up and Next Steps

at all um i really you know i'll i'll have four quick questions uh for you and the team

also and become and become an advocate um at that point um but thank you so so much david um great great presentation and hope you stay for the rest as well yeah thank you very much guys and yeah and you're the rest of the of the session absolutely

fantastic I'm soon yeah I'll do that yeah you can keep me see you guys