How we use AI

AI that works for you, and first of all for us.

We use AI every day to design, monitor and document. Here is where, how, and with which guarantees.

Talk about your AI uses Discover

The Unos approach

We only recommend what we use ourselves. AI speeds up analysis, writing and diagnosis; it never decides alone on a sensitive action.

Where we use it

Four uses, all supervised.

Design and development

Code assistants speed up the design of our themes and applications. Every change is reviewed, tested on a development environment, and only then deployed to production.

Operations and diagnosis

Log analysis, finding the cause of an incident, writing step-by-step procedures: AI suggests, an administrator validates and executes.

Documentation and reports

Meeting notes, procedures, audit reports and roadmaps are prepared with AI, then reviewed by an expert before being handed over to you.

Watch

We follow models, tools and regulation (GDPR, AI Act) so that we only recommend what is mature, and spare you passing fads.

Our safeguards

Five rules we apply to every project.

  1. A human validates

    No sensitive action (production, data, security) is launched without human validation.

  2. Backup first, test before production

    A diagnosis and a backup come before any intervention, and every change goes through a test environment.

  3. Controlled data

    We do not entrust sensitive data to an AI service without a framework: model hosted in France or run locally, or a service governed by contract and GDPR.

  4. Transparency

    We tell you where AI is involved in your project, and what remains done by hand.

  5. Measure before scaling

    A use case is tested on a small scope, its benefit is measured, then it is rolled out.

For your teams

What we build for you.

Support agents

An agent that answers your users, opens tickets and escalates to a human when needed.

Search across your documents

Ask a question in natural language and get the answer from your own documents, with the source.

Sorting and summarising

E-mails, contracts, meeting notes: sorted, summarised and ready for your teams to validate.

Workflow automation

Repetitive tasks connect your tools (Google Workspace, ERP, messaging) without re-keying.

Frequently asked questions

What we are often asked.

Is my data used to train models?

No. We choose services whose terms exclude training on your business data, and we check this before using them.

Is my project covered by the AI Act?

It depends on the use and the level of risk. At scoping, we identify with you the obligations that apply and what needs to be documented.

Can I use AI without exposing my sensitive data?

Yes: model hosted in France or run locally, data anonymisation, or a governed third-party service. The choice is made at scoping, according to how sensitive your data is.

Let’s talk about your uses.

One conversation is enough to spot the useful use cases, the data to protect and the first scope to test.

Talk about your AI uses See our AI services