
Modules built around the company's knowledge
The Data & AI platform belongs to the client: we build it in their environment, on one process or on several in parallel, with our modules, their technologies and the development their context requires. It arrives with Deployment and stays in production with Presidio.
Modules that grow with every process
We activate them to the extent the process requires: Data Foundation and Data Intelligence are often enough for a first piece of analysis, while redesigning a process around agents calls on almost all of them.
The Company Brain at the centre, the other modules around it
Agents read rules and context from the Company Brain before they act, while Data Foundation and Data Intelligence supply it with data and forecasts. Governance and Security run across every module.
The company's systems and data: they stay where they are
- 01
Data Foundation
Data ready for people and agentsPrepares the data, each item with a source, a definition and an owner - 02
Data Intelligence
Analysis and forecasts for decisionsAnalyses the data and derives forecasts and scores - 03
Company Brain
The company's knowledge, organisedKeeps rules, definitions and decisions, with the reasons behind them - 04
Agents
The work inside processesAgents carry out the work in processes, within declared limits - 05
Workspace
Where people and agents workBrings the agents' work to those who must use or approve it - 06
Governance
Rules, registers, evidenceSets the rules of use and keeps the evidence of what happens - 07
Security
Data, access and agents protectedProtects data, access and agents, within the company's security rules
X Lab: we test frontier solutions ourselves first
In X Lab, our in-house research centre, we follow what comes onto the data and artificial intelligence market and put frontier solutions to the test, to understand what they can really handle and where they stop. New releases are many and not all of them deliver what they promise, so we propose to a client only what has passed our tests. The same work keeps the people who run projects up to date: they already know what is worth using and what is not.
A sheet for every process in production
When a process goes into production we gather what the modules know about it into a single sheet, so an auditor finds in one place where the data comes from, which rules are in use and who decided.
| Voce | Esempio | Modulo |
|---|---|---|
| Dati | Anagrafiche, bilanci, andamento dei rapporti, con fonte e responsabile | Data Foundation |
| Priorità delle pratiche | Un punteggio ordina le pratiche in arrivo; versione e ultimo controllo del modello | Data Intelligence |
| Regole | Versione 14, con data e autore dell'ultima modifica | Company Brain |
| Grado di autonomia | L'agente prepara la bozza, l'analista propone la delibera | Agents |
| Dove lavorano le persone | La coda delle approvazioni nell'applicazione di istruttoria | Workspace |
| Prove e misura | 240 casi reali, soglie concordate; tempo per pratica e rilavorazioni | Governance |
| Protezioni | Identità dell'agente, dati oscurati prima dell'invio al modello | Security |
What stays with the client and what stays with us
We prefer to write it into the proposal, item by item, rather than rely on a formula. We distinguish the client's data and materials, the development done for the client, our own components that we license for use and third-party technologies, and for each we state who owns it and under what rights it is used.
Client data and materials
data, documents, definitions, rules, ontology, registers
They remain the client's, in the client's systems and in open, exportable formats
Bespoke development
agents, models and applications built for the client's processes
Built for the client; the proposal sets out ownership and terms
JoinX components
the method, Compass, generic components, reusable tests
They remain ours and we license them for use, with rights written into the proposal
Third-party technologies
models, cloud, base components
They remain the suppliers', under their licences; we choose them so that they can be replaced
An architecture you can change one piece at a time
Our approach is model-agnostic, meaning independent of individual models and technology suppliers: models, runtime environment and cloud can be replaced while rules and evidence stay intact. To connect systems and agents we use open standards, such as MCP and A2A, and we work in the client's cloud, in a European cloud or in an isolated environment. This is our idea of sovereign AI, in which the company keeps control of its data and its knowledge, decides how to govern its solutions and stays free in its technology choices.
Frequently asked questions
What is a module made of?
We build it in your environment, starting from the processes we work on, so it does not arrive as software to install. Each module brings together some of our components, which we license to you with rights written into the proposal, the technologies you already have, third-party technologies such as models and cloud, and the development your process needs. In the proposal we state, item by item, who owns each part and under what rights it is used.