Glossary: cloud, data and AI in plain words
The terms of our offers, defined in a few sentences, from the cloud foundation to AI agents. Definitions checked on 3 October 2026.
What is a landing zone?
A landing zone is the shared foundation of a cloud environment: the account structure, identities and access, connectivity, security rules and cost tracking, set up before the first applications. It is also called a cloud foundation.
Described as code, it applies the same way to every new project, on GCP, AWS, Azure or a European cloud: every team starts on safe ground, without reinventing its rules.
Going further: A secure cloud foundation.
What is SecNumCloud?
SecNumCloud is a qualification granted by ANSSI, the French national cybersecurity agency, to cloud offers that meet its security requirements. It covers a specific offer of a provider, not the whole company.
Among other things, it requires the provider and the data it hosts to be out of reach of laws outside the European Union that would let a foreign authority access them, such as the US Cloud Act. ANSSI publishes the list of qualified offers.
Checked on 3 October 2026 on the ANSSI website.
Going further: Sovereign hosting, and the article Sovereign cloud and SecNumCloud: what they change for your data.
What is FinOps?
FinOps is a way of managing cloud spending together: technical teams see what they consume, finance teams understand why, and everyone decides with full knowledge. The word joins finance and operations.
In practice, every expense is linked to a team and a project, budgets and alerts warn before overruns, and unused resources are spotted. The FinOps Foundation, part of the Linux Foundation, publishes its reference framework.
Going further: Cloud cost control.
What is platform engineering?
Platform engineering means building one platform shared by your development teams, to create, test, deliver and run their applications. A platform team maintains it and treats it as a product, whose users are the developers.
The platform offers golden paths: ready-to-use templates with quality and security already built in. The easy way becomes the safe way, and every team can focus on its product.
Going further: One delivery pipeline for all your teams, and the article Platform engineering: one shared platform for your development teams.
What is DevSecOps?
DevSecOps builds security into every step of developing and delivering software, instead of checking it at the end. The word joins development (Dev), security (Sec) and operations (Ops).
In practice, code, dependencies and images are analysed automatically on every change, secrets are removed from the code, and a risky release is stopped before it reaches your users.
Going further: Security in every release.
What is an SBOM?
An SBOM (software bill of materials) is the inventory of the components of a piece of software: the libraries it uses, their versions and their origin. When a flaw is published in a component, it tells you which software is affected.
The European Cyber Resilience Act (Regulation (EU) 2024/2847) requires manufacturers of products with digital elements, software included, to draw one up, covering at least the top-level components. Its main obligations apply from 11 December 2027.
Checked on 3 October 2026 on the European Commission website.
Going further: Security in every release.
What is RAG?
RAG (retrieval-augmented generation) has an AI model first look for the relevant passages in your documents, then answer from those passages. The AI relies on your sources rather than on its memory alone, and can cite them.
The quality of the answers depends on the documents: sources that are up to date and well described, and access limited to what each user is allowed to see.
Going further: Data ready for your AI agents.
What is agentic AI?
Agentic AI refers to AI systems, called agents, that do more than answer: they plan and carry out a series of actions in your tools to reach a goal, such as handling a request, preparing a reconciliation or fixing a test.
Because an agent acts, it needs a frame: least-privilege access, human approval of sensitive actions, a log of every action and the ability to stop it at any time.
Going further: Agentic AI under your control, and the article Agentic AI: definition, examples, and how to stay in control.
What is the AI Act?
The AI Act (Regulation (EU) 2024/1689) is the European regulation on artificial intelligence. It sorts uses of AI by level of risk: some are banned, high-risk systems must meet specific requirements, and other uses carry transparency obligations.
In force since 1 August 2024, it applies in stages: the bans and AI literacy since 2 February 2025, general-purpose AI models since 2 August 2025, most other rules since 2 August 2026. Regulation (EU) 2026/1744 moved the rules for high-risk systems to 2 December 2027, and to 2 August 2028 for those built into products.
Checked on 3 October 2026 on the European Commission website.
Going further: AI agents under control, and the article Securing AI agents: prompt injection, shadow AI and the AI Act.