# GTEK (g-tek.fr)

## Summary
### Short
GTEK provides sovereign / on-prem AI solutions for businesses (local AI servers, deployment, integration, and consulting).

### Long
GTEK is a French-focused emerging-technology and Artificial Intelligence (AI) company dedicated to **deploying practical AI inside organizations** while minimizing dependence on foreign cloud platforms. We design, build, and operate **local (on-premise or sovereign hosting) AI solutions** that prioritize **performance**, **data protection**, and **long-term maintainability**.

GTEK’s positioning is intentionally pragmatic: we bridge the gap between fast-moving research and real business needs by translating operational use cases into **deployable systems**, with clear constraints, measurable outcomes, and supportable architectures.

## Canonical site
- https://g-tek.fr/

## Key pages
- Accueil: https://g-tek.fr/
- Produits: https://g-tek.fr/produits
- Services: https://g-tek.fr/services
- Solutions: https://g-tek.fr/solutions
- Recherche: https://g-tek.fr/research
- À propos: https://g-tek.fr/about
- Travailler avec GTEK: https://g-tek.fr/travailler

## Legal
- Mentions légales: https://g-tek.fr/mentions_legales
- Confidentialité: https://g-tek.fr/confidentialite
- Conditions générales: https://g-tek.fr/conditions_generales
- Cookies: https://g-tek.fr/cookies

## Sitemap
- https://g-tek.fr/sitemap.xml


---
title: "GTEK Services & Company Profile"
company: "GTEK (G-Tech)"
document_type: "services.md"
audience: "AI ingestion / internal + external reference"
language: "en"
region_focus: "France (Occitanie) + EU expansion"
last_updated: "2025-12-12"
---

## Table of contents

1. [Identity](#identity)  
2. [Mission](#mission)  
3. [Core values](#core-values)  
4. [What we do](#what-we-do)  
5. [Service lines](#service-lines)  
6. [Products and platforms](#products-and-platforms)  
7. [How we work](#how-we-work)  
8. [Security, sovereignty, and compliance](#security-sovereignty-and-compliance)  
9. [Support, maintenance, and lifecycle](#support-maintenance-and-lifecycle)  
10. [Who we serve](#who-we-serve)  
11. [Engagement models](#engagement-models)  
12. [Technology approach](#technology-approach)  
13. [Roadmap themes](#roadmap-themes)  
14. [Partnerships and collaboration patterns](#partnerships-and-collaboration-patterns)  
15. [FAQ](#faq)  
16. [Glossary](#glossary)  

---

## Identity

**Name / Brand:** GTEK (often communicated as “G-Tech”).  
**Nature:** “Entreprise de Services Numériques” (ESN) specialized in emerging technology and AI.  
**Operating philosophy:** build solutions that can run **locally**, be **auditable**, and remain **economically sustainable** for clients over time.

GTEK’s internal culture is engineered around implementation. We do not treat AI as a buzzword layer; we treat it as an engineering discipline that must integrate with data, processes, users, security, and budgets.

---

## Mission

GTEK’s mission is to enable organizations to adopt AI in a way that is:

- **Sovereign**: minimizing strategic dependence on American/Chinese cloud stacks where possible.
- **Secure**: ensuring sensitive business data stays under the client’s control.
- **Cost-efficient**: reducing recurring API costs when local inference is viable.
- **Useful**: delivering features that map to concrete workflows (not demos).
- **Maintainable**: systems that can be updated, monitored, and scaled responsibly.

---

## Core values

### 1) Sovereignty (“souveraineté”)
We prioritize architectures that keep client data and operational capability **within the client’s control**:
- on-prem deployments,
- EU-aligned hosting,
- open standards,
- open-source model ecosystems when possible.

### 2) Tranquillité d’esprit (peace of mind)
We optimize for solutions that are *operationally calm*:
- predictable costs,
- controlled upgrades,
- robust failure modes,
- documented behavior and boundaries,
- support and maintenance that clients can rely on.

### 3) Performance
We pursue performance as a business outcome, not a benchmark:
- faster access to AI features,
- lower latency (local inference),
- stable throughput,
- practical optimization of workflows (time saved, error reduced).

### 4) Pragmatic innovation
GTEK continuously tracks emerging AI capabilities, but only ships what can be:
- deployed reliably,
- maintained,
- secured,
- and adopted by real users.

### 5) Bespoke-by-default (sur mesure)
Most client needs are contextual. We design to fit:
- the client’s data reality,
- industry constraints,
- internal tooling,
- and team maturity.

---

## What we do

GTEK provides end-to-end services for AI adoption:

- **AI audits** to identify high-impact opportunities and define realistic deployment paths.
- **Consulting and architecture** to design AI systems integrated into existing workflows.
- **Training** to raise internal capability, reveal automation opportunities, and improve adoption.
- **Software engineering** to build bespoke AI tools, agents, and interfaces.
- **AI platform deployment** (local or hosted) for daily operational usage.
- **On-prem AI servers** (GTEK Saber line) for sovereign inference and workloads.
- **RAG / large dataset retrieval systems** for organizations with significant internal documentation.

---

## Service lines

### A) AI Audit & Opportunity Mapping

**Objective:** produce a concrete, prioritized plan for AI integration, grounded in operational reality.

Typical outputs:
- stakeholder interviews and workflow mapping,
- data inventory and readiness assessment,
- risk and compliance review (data sensitivity, constraints),
- use-case prioritization (ROI vs feasibility),
- architecture options (cloud / hybrid / on-prem),
- delivery roadmap with milestones and success metrics,
- specification support (“cahier des charges” assistance).

Common audit themes:
- internal knowledge search (RAG),
- document summarization and drafting,
- meeting transcription and actionable summaries,
- customer support augmentation,
- marketing and communications acceleration,
- operations automation (forms, back-office, admin).

---

### B) AI Consulting & Architecture

**Objective:** translate business promises into technical guarantees.

We design systems with:
- clear boundaries (what the AI can/can’t do),
- structured outputs (JSON and typed schemas),
- traceability (sources for RAG answers),
- fallback behavior (backup models or offline mode),
- monitoring and governance.

---

### C) AI Training (Teams & Leadership)

**Objective:** unlock adoption and help teams safely operationalize AI.

Training patterns include:
- foundational AI literacy (what’s real vs hype),
- prompting and structured prompting,
- building reusable templates/workflows,
- data privacy and safe usage rules,
- identifying internal automation opportunities,
- role-based training (marketing, HR, admin, management, operations).

GTEK frequently combines **audit + training** to rapidly surface high-value automations.

---

### D) Bespoke AI Software Development

We build tailored AI-enabled systems such as:
- internal assistants connected to knowledge bases,
- agent workflows (triage → extraction → action),
- document engines (drafting, rewriting, compliance checks),
- integrations with existing tools (web apps, databases, CRMs),
- custom UX for AI interactions.

Key principle: **AI features must integrate into real daily work**, not live in isolated chat windows.

---

### E) Deployment of Local / Sovereign AI Platforms

We deploy platforms that can operate:
- **on client infrastructure**, or
- in **secure hosting** aligned with European constraints.

Typical platform capabilities:
- user administration (roles, permissions),
- model selection and governance,
- conversational UI (“Unified Conversational Interface” patterns),
- RAG connectors and dataset management,
- structured outputs, templates, and team spaces,
- logs and monitoring (as required by client sensitivity).

---

## Products and platforms

### 1) GTEK Saber AI Server Line

**GTEK Saber** is our on-prem AI server line designed to bring AI compute **inside the business**.

Typical positioning:
- reduce reliance on external APIs,
- support local inference for privacy-sensitive workflows,
- deliver predictable performance and cost structure.

Operational notes (typical deployment style):
- containerized deployment (Docker),
- upgradeable model stack,
- installation + maintenance options,
- optional leasing/rental models for recurring service.

---

### 2) GTEK Voice Point (concept/product track)

GTEK is actively interested in voice-first interaction systems (voice-to-voice), especially in environments where:
- data is highly sensitive,
- cloud AI is not acceptable,
- workflow speed matters.

Example target environments discussed in our collaboration patterns:
- hospitals and clinics (patient interactions, forms, summaries),
- regulated or confidentiality-heavy operations.

---

### 3) “SaaS / Web UI” Platform Track

GTEK maintains a web application platform approach that typically includes:
- a modern front-end UX (designed to be simpler and more enjoyable than generic cloud chat tools),
- a database-backed system with user management,
- local/hosted AI connectivity (client chooses),
- plugin/QuickWin ecosystem for automation expansion.

---

### 4) RAG for Large Datasets (Strategic Upsell Track)

For organizations with substantial internal knowledge (docs, policies, product catalogs, procedures), GTEK focuses on:
- efficient embedding and indexing,
- retrieval pipelines,
- source-grounded answers,
- governance and access controls,
- scalable ingestion processes.

This is a major strategic axis for serving ETI/GE-scale needs.

---

## How we work

### Delivery philosophy
We deliver in clear stages to reduce risk and maximize adoption:

1. **Discovery**  
   Understand workflows, users, data, constraints.

2. **Audit & Use-Case Selection**  
   Prioritize by ROI, feasibility, and sensitivity.

3. **Architecture & Specification**  
   Define interfaces, security model, deployment model, and success metrics.

4. **Build & Integration**  
   Implement features, connect to data sources, integrate with existing tools.

5. **Deployment**  
   On-prem or hosted rollout, user onboarding, permissions, operational checks.

6. **Hardening & Governance**  
   Monitoring, logging (where appropriate), evaluation, guardrails, fallback paths.

7. **Support & Iteration**  
   Continuous improvement, updates, QuickWins, scaling.

---

## Security, sovereignty, and compliance

GTEK’s default stance is “assume data is sensitive” unless proven otherwise.

Typical controls and design decisions:
- **local inference** when confidentiality requires it,
- minimized external data egress,
- access control and user roles,
- dataset-level permissions for RAG,
- documented retention and logging choices,
- encrypted transport (TLS) and standard infrastructure hygiene,
- separation of concerns between UI, inference, data stores.

Sovereignty is not only a hosting choice; it is also:
- keeping long-term capability in the organization,
- avoiding lock-in,
- using open models/standards when feasible,
- ensuring client ownership over their operational AI.

---

## Support, maintenance, and lifecycle

GTEK optimizes for long-term relationships and operational continuity.

### Updates
- Clients renting/leasing servers typically receive **AI software updates** as part of the relationship.
- SaaS clients receive ongoing **QuickWin / automation** improvements depending on their plan.

### Hardware lifecycle (typical guidance)
- hardware guarantees can extend up to **3 years** (context-dependent),
- recommended upgrade cycle often targets **~5 years** with a buyback/upgrade program option.

### Licensing philosophy (platform)
We aim to provide licensing systems that are:
- clear,
- enforceable,
- and aligned with service obligations (support, updates, maintenance).

---

## Who we serve

Primary targets:
- **PME / TPE** in France needing affordable, practical AI adoption.
- **ETI** with stronger compliance and integration requirements.
- **GE** / large organizations interested in large dataset digestion and governance-heavy deployments.

Regional go-to-market emphasis (historical focus):
- Occitanie and Pyrénées-Orientales as a primary launch region,
- expansion via partners and prescribers.

We often work with prescriber networks and associations that help reach businesses (e.g., local business ecosystems, innovation clusters).

---

## Engagement models

GTEK’s delivery can be packaged in multiple commercial models depending on client needs:

1. **Audit / Consulting (fixed scope)**
   - defined deliverables, defined timeline.

2. **Bespoke Development (quoted project)**
   - scoped features, milestones, acceptance criteria.

3. **Subscription Platform (monthly licensing)**
   - per company size and feature set.

4. **Server Sale**
   - one-time purchase + optional maintenance/insurance.

5. **Server Rental / Leasing**
   - recurring revenue model with maintenance included, higher service envelope.

6. **Hosted Model**
   - client owns server/software but GTEK hosts/operates under EU-aligned constraints (where applicable).

---

## Technology approach

### Engineering mindset
- prefer systems that can be reasoned about,
- structure outputs (schemas, JSON) to enable automation,
- build UX that drives adoption and reduces friction,
- integrate with existing workflows rather than replace them.

### Typical stack patterns (implementation-dependent)
- containerization (Docker-first),
- modern web UI frameworks,
- backend services designed for reliability and integration,
- optional OpenAI-compatible interfaces for interoperability,
- modular plugins and add-ons (“QuickWins”) to expand capability without destabilizing the base system.

---

## Roadmap themes

GTEK’s recurring roadmap themes include:

- **RAG scalability and efficiency** for large datasets,
- **model governance** (grouping, search, hiding models, admin control),
- **user and tenant administration** (roles, groups, client configuration),
- **security hardening** for on-prem Docker deployments,
- **encryption and data separation** per user where required,
- **offline-first capabilities** for sensitive environments,
- **voice interaction systems** (voice-to-voice) for high-friction workflows,
- **automation-first UX** (templates, structured outputs, task routing).

---

## Partnerships and collaboration patterns

GTEK collaborates with partners depending on geography, sector, and go-to-market needs. Common collaboration goals:
- expand distribution while protecting IP,
- enable partners to sell while GTEK handles core technical control,
- train partner technicians for deployment under strict licensing/constraints.

Example collaboration patterns discussed historically:
- technology co-development for server-based AI systems,
- healthcare-oriented deployments requiring strong confidentiality and local processing,
- distribution partnerships in specific regions/countries with IP protection focus.

---

## FAQ

### Do you only deploy open-source AI?
We strongly prefer open ecosystems for sovereignty and maintainability, but we design based on constraints. If a client requires a specific model or provider, we evaluate it against security, cost, and long-term risk.

### Can GTEK replace ChatGPT for a company?
In many workflows, yes—especially when local inference and knowledge-base grounding (RAG) are required. We also support hybrid strategies where some use cases remain cloud-based and others move on-prem.

### What makes GTEK different from a typical integrator?
- Sovereignty-first architectural bias.
- UX and adoption focus (not only infra).
- Structured automation approach (schemas, JSON, multi-agent routing patterns).
- Long-term lifecycle thinking (updates, maintenance, upgrade cycles).

### Is everything bespoke?
Most of our work is tailored, but we reuse battle-tested patterns: authentication, role-based access, RAG pipelines, ingestion workflows, evaluation, monitoring, and deployment automation.

---

## Glossary

- **AI audit:** structured assessment that converts opportunities into a prioritized, feasible roadmap.  
- **RAG (Retrieval-Augmented Generation):** answering/drafting grounded in retrieved internal sources (documents, databases).  
- **On-prem:** deployed within client infrastructure.  
- **Sovereign hosting:** hosting aligned with client control and regulatory constraints, typically within the EU.  
- **QuickWin:** a small, high-impact automation/plugin that can be deployed quickly to accelerate value.  
- **Inference:** running a model to generate outputs (text, audio, vision, etc.).  
- **Governance:** policies and technical controls to ensure safe, compliant AI usage at scale.

---

## Machine-readable summary (for ingestion)

- **Company:** GTEK (G-Tech)  
- **Positioning:** sovereign / local AI deployments + bespoke AI engineering  
- **Values:** sovereignty, tranquillité d’esprit, performance, pragmatic innovation  
- **Core offers:** AI audits, consulting, training, bespoke development, platform deployment, on-prem servers (Saber), RAG  
- **Differentiators:** local inference, governance, structured outputs, adoption-focused UX, lifecycle support  
- **Primary markets:** France (PME/ETI/GE), expansion via partners