Introduction

Good evening, I'm very sorry not being able to be there in person. So tonight I'll be speaking about Pupao, our AI agent platform, and I'll show you quickly our vision and I'll show you the product, our key point and the problem we solved.

Why AI Sovereignty and Governance Matter in Europe

We know that as a European, we are the weak link in the AI race because we at the moment we have not that much investment in models we don't have that much investment in the infrastructure so we definitely need to be on that race because we know that AI will be in our next future the the

driver for industry company corporate small businesses for everything so we we we decided to build a platform that let us be able to govern, manage and be able to use safely with our value, with our objective.

Where Pupao Runs: SaaS, On-Prem, and Edge Deployment

Let's start from where our platform can be used. You can use our platform as a product as software as a service through the web using our public web accessible solution you can use our platform within your data center so you can just use

within your data center but you can even use the platform on the edge on your desk with a small ai powered machine that you can freely buy on the market there are plenty of them so you can and use an AI -powered machine that runs and lets you use our platform.

Mobile, Plugins, and Simple App Integration

Our platform is available on mobile platforms like Android, iOS. You have a WordPress plugin.

You have the ability to use our agent, to use our solutions through your own application with the right two lines of code.

Core Promise: Governed AI Access Across the Organization

And the main thing that you can do with our platform, it's access different shape and different level of AI within your organization. So what we give to our customers is the ability to have a governance of the AI while using AI itself.

Keeping Data, Metadata, and Processes Private

Today, if you're using AI, if you're using any hyperscaler solution, you are basically giving your data with all your information, with all this kind of data that you have to other providers. Of course, there are contracts, of course, there are GDPR, there are rules, but other than your data, what you're sharing with the providers, with external providers, are your metadata, are your habits, your processes, which are not subject to the same policy that your data is. So what we do is to give you the same thing that the other providers give, but in a very, very private and secure way.

So our platform gives you the ability to use agents. Agents mean AI that can do stuff for you, that they can prepare documents, that can do billing, that they can do invoicing in your organization, without the hassle to give data outside. side.

Agents for Real Work—and a Model-Agnostic Approach

The AI itself in the platform, it's a commodity. It's something that you just configure with a click.

We are not sponsoring any particular AI. We work with any AI.

We work with frontier model which is GPT, Cloud, whatever you name, we use it.

Open-Source Models Run Locally

But we mostly focus on open source model that you can run locally.

So, since we want that you have your data and your AI locally, we put a lot of effort

all the instruments that let you run powerful AIs in your data center, in your computer, in your laptop, in your desktop.

Governance-First Architecture

The approach we give is to the total managing of the AI and what you do. 1So we started from the ground up to put governance as the main point for you to control. So you have the ability to basically check, control, and decide what data in your platform can go, where this data go, what AI you're using, how you're using the AI, and what's the level of compliance for each interaction you have with your AI.

No-Code Setup for Non-Developers

Moreover, we have a no -code approach that gives users the ability to configure everything they want of their AI without being a programmer. The platform gives the ability to know technical people, know technical persons, to in a very easy way be able to understand and do many, many technical stuff.

Knowledge Bases via Drag-and-Drop

Like an example, put a knowledge base to an AI. Today, if you want to put a knowledge base in AI, it should be and is, in most cases, as easy as just drag and drop a document. We do exactly this.

Granular Access, Retention, and Update Controls

You can have your agent, give information to your agent, manage the access of this information within your organization in order to be able to decide who in your organization has access to each piece of information, You have the ability to decide for how long those information are accessible to your user or your customer, depending on the use case. And you can even decide how frequently those information can be updated.

Our platform gives you the ability to do very, very, very complex stuff without being a developer.

Connectors: Messaging, Email, and Voice

1You can set a connector like Telegram, email, WhatsApp.

Soon we are enabling VoIP, Voice over IP, so Voice as an input connector for your agent. So in a very easy way, you will be able to just let your agent talk with you

or with your customer, with your employee, in a very easy, very convenient way without the need for any integration, for any work on the development side.

Tools and Data Connections Without Custom Code

Same thing, it's our approach to tools. Tools are what make AI today great. Tools are what AI use in order to accomplish objective.

In example, if you want to let an AI manage your customer base, maybe you want to give an access to a specific database, to a specific data to an AI,

today you usually need to code that part with our platform you can basically just set the parameter of the database or whatever you want to input to give to the AI and you have access to that so it's very easy to configure to use and to and to work with the the adoption

Adoption and Use Cases in the Field

of the solution at the moment it's in very different span of sectors we have

an example in Italy we have been a chain supermarket chain with 80 plus locations in Italy that are using our solution in order to help employees to access the the vast information that the HR have for them, for the different employees. So the employee of this chain interact with the HR through our solution to a specific agent, exactly, and helping them for that.

Other companies are using our solution, in example, to manage the incoming invoice. Those invoices need to be categorized, need to be approved, need to be reviewed against, in example, for example, Excel file against Google Drive document, etc. And they now have an agent integrated with their ERP solution, their Zucchetti, and they just sold, they avoid four hours work per day in a very easy and convenient way. Those are just some examples.

In Torino, there are Fondazione Charity, they are using our solution in order to have an AI that is sovereign, that does not use any external AI models other than the ones they have in -house in order to be able to manage personal data, the financial document in a very private way.

Platform Modules: Data, Agents, and Automation

The platform itself gives you the ability to use agents, gives you the ability to handle knowledge bases, So big data and work with those data, ask those data and extract information, graphical, whatever you need with those data.

Marketplace of Ready-to-Use Agents

We have, of course, generation. We have a marketplace that gives you the ability to have a ready -to -use agents in a very specific field. So you don't need to think about what kind of agent you want.

Scheduled Tasks and Personal Automation

We have the ability to set tasks. So you can ask your AI to set a task that every Monday you do the brand positioning analysis for your brand and send that to you by mail every day.

I use it to get the meteo, to get what to do with my daughters in the weekend for events and whatever. But of course, you have many use cases, just fantasies telling it.

Visual Agentic Workflow Builder

Plus, we have a very powerful agentic workflow builder that lets you, in a very graphical way, build a very complex workflow that you can use to do any task you can think about.

In this case, I'm just showing you a very simple workflow that queries some repository repository of information, authenticate, use this authentication to get some more article,

get those article and then use different AI models to do a stamp, to do the review of the news of the day and then put all those news in an API. So it's a very specific use case.

Usually you do these, those kinds of use case by coding the use case. In this case, you just put these blocks, configure these blocks, and say to the different parts of AI what they need to do.

Auditability, Policies, and Compliance Controls

Again, everything you do in the platform, it's recorded in order to give you the ability to understand how your users use AI and if they do in a compliant way.

You have the ability to set some policy. Those policies choose what users can do, what kind of model they use, if the model needs to be in the European Union, they can be whatever, if you want to be absolutely sure about your data.

Conclusion

So we are a leader on the AI that gives you control on how AI is used in your company and gives you the ability as well to get the most of the AI without being a technician.