Solutions

Sovereignty & security

Deploy AI while retaining control of your data, access and hosting. External model APIs are used only at your request.

Summary

Keep a clear view of your system.

Runs on your premises or in a cloud chosen with you, such as OVH. A model’s origin does not determine where it runs.

Your choicesExecution location, access and any use of an external API are agreed with you.
LinkTecSystem design, configuration and deployment, documented for each project.
Identified third partiesA telephone operator, model providers or a hosting provider may be involved depending on the configuration; they are identified.
Your AI systems
Hosted on your premises or in a cloud chosen with you. The environment, access and components are documented for each project.
Models
Models from various origins, selected for your use case. Licences, versions and commercial rights are checked before any specific commitment. No external model API without your request.
Handover
Deliverables, rights and handover conditions are set out in the proposal, taking licences into account.

Together, we define safeguards, operational responsibilities and handover conditions for each system.

Confidentiality, NDA & data

Your data protected by an explicit framework.

From the first exchange of data to project completion, LinkTec defines with you the scope of use, authorised people and handover arrangements.

Request an NDA

A confidentiality agreement before any project data is shared. Your own template can serve as the basis.

Data path

  1. Non-disclosure agreement (NDA)

    A non-disclosure agreement is signed before the first exchange of project data, including during initial discussions and before a quote. It covers the information shared, its use and the obligations of contributors. We can work from your own template.

  2. Access for authorised people

    Contributors access only the data required for their role in the engagement. Access scope and confidentiality obligations are formalised before they get involved. Data exchange arrangements and the hosting environment are selected with you.

  3. Data use and AI models

    Data entrusted to us is used within the agreed scope of your project. Where model providers are involved, non-use of your data to train their models is a contractual selection criterion: their commitments are checked and attached to the agreement. Adapting a model for your own use is covered separately within the scope of your engagement.

  4. Personal data & processing agreement

    Where an engagement involves personal data, a data processing agreement (DPA) specifies the purposes, retention periods, parties involved and handling of data at the end of the contract. Service providers are identified; you are informed in advance of any additional subcontractor involved in the engagement.

  5. Retention planned from the outset

    Retention periods for project data and end-of-engagement arrangements are agreed in the contractual framework. We distinguish data needed for the project, working copies and documents subject to retention requirements, such as accounting records.

  6. Return, deletion & written confirmation

    Deliverables and source data are returned in open, reusable formats. Exit arrangements also cover deletion of working copies, indexes and exports under the agreed terms, with written confirmation of deletion available on request. Records subject to retention obligations follow their applicable requirements.

Delivered systems

Clients’ AI systems

This section covers systems delivered for your projects: AI models and agents, AI infrastructure, simulation and decision support, and Voice AI. Every point is agreed with you and documented for your system.

Hosting

Your systems run on your premises or in a cloud chosen with you, such as OVH, without restriction to one provider or region.

  • AI infrastructure: server design, sourcing and assembly, GPU configuration, networking, model deployment, monitoring and maintenance.
  • Service levels, including monitoring, maintenance and response times, are agreed in writing for each engagement.

Access

Together, we define roles, administration, data access and activity records, as described in the system documentation. Logging, retention, encryption and backups are agreed for each system according to the chosen environment.

Dependencies

  • Instructions, RAG, configuration and fine-tuning are among the techniques we use in system design.
  • French, American or Chinese models may be used; a model’s origin does not determine where it runs.
  • Licences, versions and commercial rights checked for every selected model.
  • Versions, data flows, telemetry and third-party components are documented. Every required third-party component is identified, including for on-premises execution.
  • No external model API without your request.

Takeover conditions

Deliverables, formats, documentation, rights and handover conditions are set out in the proposal; third-party models and components remain subject to their own licences.

Clear follow-up if an incident occurs

If a confirmed breach affects project data, the client is informed without undue delay of its nature, scope and immediate measures. A written report is provided when the incident is closed. Contacts and follow-up arrangements are specified for the engagement. To report an issue, write to contact@bn-linktec.frwith “Security” in the subject line.

Data processed by linktec.fr and its client portal is described in the privacy policy.

Next step

Let’s examine your deployment requirements.

We review with you the hosting, access, dependencies and handover arrangements of your system.