Skip to content

Service · Our specialty

AI & Automation

Our specialty: private LLMs adapted to your data and hosted on your premises, with automations running on them. When your data allows it, we also build your solutions on market models (Claude, GPT…), through API requests.

  • An AI adapted to your business vocabulary and tasks
  • Your choice of model: private on your premises, or a market model via API
  • Goals set together, with the way to measure them
Book a free call

30 minutes · no commitment · video or phone

Our reviews on Trustpilot
Call us

Pre-selected topic: « AI & Automation »

  • Private AI

    An open-source model (Mistral, LLaMA, Qwen or Gemma) adapted to your business, installed on your premises or on our servers, in France or in your country, with no external AI API.

  • AI via API

    At your request, we integrate Claude, GPT or other models through their APIs: for example if you don’t have the hardware (GPU) to run a fine-tuned open-source model, or to start without a server. Requests are then processed by the model’s provider, and it is written down during scoping.

  • Sourced answers

    The documents used for an answer are displayed, so your teams can check.

01 · Scope

Defined before we start and carried over unchanged into the proposal, with nothing added mid-project.

  1. 01Private LLM fine-tuned on your examples: vocabulary, answer formats, recurring tasks
  2. 02Solutions on market models (Claude, GPT…) through API requests, at your request, for example with no GPU server
  3. 03Search across your documents (PDF, Word, CSV, text) with sources displayed
  4. 04Installation on your premises or on our servers, in France or in your country, depending on the project
  5. 05Web interface, accounts and roles, export of answers to PDF and Word
  6. 06Phone AI for booking appointments and taking requests
  7. 07Process automation: data entry, extraction, matching, sending
  8. 08Connection to your existing tools (CRM, ERP, intranet), defined during scoping
  9. 09Computer vision, sensors and embedded AI

02 · How it runs

  1. Scoping

    Target tasks, data, choice of model (private or API), hardware and acceptance criteria, in writing.

  2. Adaptation

    Data prepared with you, then the model or the requests are adapted.

  3. Evaluation

    The solution is evaluated on your cases. If it fails, it does not go into production.

  4. Go-live

    Installation, health checks, operations documentation and training.

FAQ

30 minutes to understand your need, before any proposal.

Book a free call

30 minutes · no commitment · video or phone

Our reviews on Trustpilot
+33 7 83 83 94 81
Book a free call