Introduction

Hello everyone, so this is the demo of Acoree. Acoree is an AI startup, of course.

The Problem: Context Depth and Knowledge Silos

We are talking about AI here and what we are solving is the context depth that of course when you have to switch from Cloud to JGPT to Gemini

or you want to share something with your colleagues, with your teammates and their Claude, for instance, doesn't know anything about you and about what you have done in the last months, for instance.

The Company Brain: A Shared Knowledge Base

So we created this platform for companies, mostly for small and medium enterprises, to have one company brain shared among all the teammates, all the people inside the company

and we start seeing this platform from the knowledge base so this is the company brain

Uploading and Training on Any Content

you can upload in this environment anything from different sources

so you can upload PDFs videos, podcasts reels even from Instagram and any other social media and here you see even cloud skills so it's everything is

compatible with a close skills but the most important thing now to highlight is that you can upload anything from okay from your computer yeah okay and what it does is to be uploaded here and to be trained so this training happens for you

Reuse Across the Organization (and Token Savings)

and for the whole company that's why we save so many tokens because your teammates don't have to re -upload the same document every time anyone in company has to use that specific contract for instance document so one upload for the whole organization and this is the brain that is the foundation

From Knowledge to Action: Creating AI Agents

once we build the knowledge the brain what we can do is to create agents so think of a of an agent here that we call echo as a project include so you can

have even folders of agents but let's start for instance with HR and people so you see you have the echo brain of that specific agent you are the items from your knowledge in the in this demo it's already uploaded and you can customize the settings of the specific agent.

Customizing Agent Behavior and Capabilities

So you can change the voice because you can talk to the agent. You can customize the instructions just as in custom GPTs or in cloud projects.

You can give the agent more capabilities like you can go or do not go to the internet. you can execute or cannot execute code you can link specific skills and so on so forth you can even connect any mcp you can think of so let's start with uh with an example i'm here in uh in

HR Example: Requesting Time Off via an Agent

the HR echo and I ask how do I lock time off for my child's illness so instead of going to the HR manager I ask the the echo for these and the echo replies with the knowledge it has in the in the knowledge and then I can I can ask feel fill the form on my behalf.

And what it does is, in an agentic way, to fill the form. Obviously, it needs all the data. And then what it can do is to send that form to the HR manager on your behalf. So it's a full agentic tool.

A Unified Workspace for Everyday Work

Now, another example I can think of is if we go, okay, this is the home page, so you can customize the home page as much as you want, it's all HTML, so here we made some fake metrics that you can share with your colleagues and they go right here.

Marketing Example: Generating a LinkedIn Post and Image

but if you go for instance to marketing and contents and you ask write the linkedin post for the launch of flows that in this demo is one of the product the the eco will write the post on your behalf in a few seconds of course and then you can even ask create an image for this post so you have the intelligence of the of top tier LLMs like in this case it's using cloud for this part and now it's using chat GPT to create an image something something you cannot do in Cloud.

So in one environment, you have all the tools you can think of and you have ever used with other LLMs, and everything is shared among all the

So, in this case, I didn't give any, the prompt was very simple. So, here it is. What else?

We have one more tool. I would like to, two more tools. Yeah. I would like to highlight.

The Super Agent: One Personal Avatar Across All Knowledge

One is the super agent. so right now when you use how many of you use cloud okay how many of you use chat gpt gemini how many of you use all of them okay so uh when you use a custom gpt or cloud project

The problem is that the cloud or GPT or Gemini, say, knows everything about that specific project, but doesn't know anything about all the other knowledge outside that specific project. So we are solving this problem, these silos we are building inside our companies with a super agent. So think of it as your personal avatar.

What the Super Agent Can Do

So the super echo or super agent in this case knows everything you know, has access to everything you have access to and can do everything you can do. So if you can create a new document in a specific folder, as a person, as a human, your avatar, your AI agent, can do the same.

Example: Checking Grant Eligibility Using Multiple Domains

So in this case, I'm asking the agent to summarize whether I can access a grant or not. And I'm uploading the, let's see if I can find it. the document and in this case this super agent is um is using the knowledge from

okay uh from different uh uh echoes so it's using the knowledge from uh the the echo that deals with tenders the echo that deals with administration administration, with economics to see whether the financials you must have to apply to this standard is okay with you.

And so we are solving these silos in this way. So it takes obviously a few seconds, but eventually we'll have the answer.

in this case you see there's an error but that's completely normal because it's writing the code for itself so it's trying to solve problems that humans cause in this case I didn't give the the agent the the access to the the folders

So he's giving himself the access. We'll come back to this in a few seconds.

LMS Feature: Building Courses from Company Knowledge

Before this, I would like to share with you one more feature that is the LMS, so Learning Management System. So once you have all the knowledge uploaded, what you can do is to create a new course for your organization.

In this case, I can say, okay, onboarding. Nice. I was about to create it before joining this demo. So we're doing it live.

Course Creation Wizard and Structure

And you see here is everything is driven. There's a wizard. so I can tell it okay create this course in English use a friendly tone of voice

and it must be for beginners it's an onboarding afterwards after all and here I can say it's Acme products and these new hires and here two chapters with with one quiz at the end of each chapter. So what it does is to create.

OK, here I have to say what to use, the sources. Here we are. And what it does now is to create the outline.

Human-in-the-Loop Review and Final Generation

Of course, the human, as we say, is in the loop. so as as a person in charge I can change anything I don't like what I want to to improve so I can create a new chapter I can change the the names I can change the questions of the quiz I can do anything i want actually so here we are i can i can move the chapters i can add anything here

a podcast for instance would be nice i have to say okay use this specific knowledge so that that the AI knows what's the basis, the info it has to use, and then I generate the course. So here we are.

We have everything in one place, and we just say Generate, and it generates everything. So now I will not generate everything, otherwise it takes too much time okay let's see okay in the meantime I go back

to the the super echo that is here and I should have the document ready let me me see yes here we are so what it did was to take the PDF and put all the all the data inside so what it did was to turn the PDF into a word so that it can edit the file and then taking all the data from the knowledge it has filled

the document up so you have everything ready to be shared in a couple of minutes instead of having to manually look for information and uploading filling the form up okay I have 30 seconds more I will take these few

Governance: Permissions, Groups, and Onboarding

seconds just to say that everything is for companies so we know governments governance and permissions are fundamental and so what we do what we did is to create this organization based on your organization so you can say

create a group that is sales okay sales and then you can onboard you can you can can add members to this specific group or other groups from an Excel, from your, for instance, factorial, and any other HR tool you have.

Conclusion

So this is Acoree.

Maybe later I'll show you the course if you have time and want to do so.

OK.

Thank you very much. Thank you, Lorenzo.