Security and compliance
What this page covers
The certified management system behind the work, how security enters the build, what happens to data, and what a supplier qualification can ask for and get.
Security here is a way of working, not a service on sale. The management system is certified to ISO/IEC 27001, and the practices under it are written, each tied to a control and to the evidence that proves it ran. The page answers in the order a supplier qualification asks.
Certified, with a scope written down
The information security management system is certified to ISO/IEC 27001:2022 by Bureau Veritas, under certificate IT356335, for consulting and the development of software applications. Under it sit written procedures, each naming the control it answers to and the evidence it leaves. Access rights are reviewed twice a year, with the exports attached to a ticket and the outcome approved by each system's owner. Hiring checks and supplier clauses sit under the same system.
Go deeper
Security inside the build
Secure development follows the OWASP practices, and the checks that enforce them run in the pipeline, not in a review at the end. The AI tools that write code are a separate subject.
Checks that stop a release
Every build is scanned for flaws in the code, in its libraries and in the container image. A critical finding fails the pipeline, and only signed images reach production.
Secrets out of the code
Keys and passwords live in a vault and are injected at build time. The values left in the repositories are decoys, so a leaked file is not a leaked credential.
Attacks rehearsed in a lab
Vulnerability classes affecting systems we run are reproduced in an isolated lab, an unpatched build beside the patched one. When the proof fails on both, the outcome is inconclusive, never a pass.
Events collected, and acted on
Security logs from our systems and from those we run for other companies flow into monitoring that flags unusual access, beside dependency alerts and public advisory feeds read on a set cadence. Each system also carries a monthly check of updates and security, with its own ticket and its own test. When a vulnerability lands on a component many systems share, one coordinated action covers them all. Incidents follow a written procedure: detection, triage, containment, remediation, notice to the company affected and a review once it is closed.
Go deeper
Where data sits, and when it goes
A security questionnaire asks three things about data: where it is, who reaches it, and what happens at the end of the contract.
A region, and encryption either way
Hosting sits in the European Union by default. Data is encrypted at rest with AES-256 and in transit over TLS 1.2 or later, with keys in the provider's key management service.
Sign-in on the company's own identity provider
Access runs through SAML single sign-on on the identity provider of the company that buys, so passwords and the second factor stay on its side.
Export first, deletion with a record
At the end of a contract data leaves in open formats, then goes from production and from the backups within an agreed window, with a record of what was cleared.
Go deeper
AI in what we deliver
Where a delivered system contains AI, the first question is our role under the AI Act, provider or deployer, settled system by system with the risk level and the documentation that level requires: model cards, a risk assessment and the record of human oversight. Models run in enterprise-managed environments in the European Union, under contractual no-training and no-retention terms; vision models trained in-house run on our own infrastructure. Every AI output reaches a person before it is committed, carries an AI-assisted label, and is logged with the action taken on it.
Go deeper
What a supplier qualification gets
Enterprise buyers qualify a supplier before they sign. The work of answering is done once and kept, so the next questionnaire starts from the last.
Questionnaires answered, documents ready
A ninety-three-question enterprise security questionnaire has been answered in full and annexed to an order form as a security exhibit. Certificate and scope, the policy set and the data processing agreement are issued on request.
Policies from the other side
Companies bring their own supplier security requirements and AI policies: which tools may touch their code and data, and where that data may go. The stricter rule wins.
Accessibility built in, then checked
What we build is designed against WCAG 2.1 AA from the first screen and checked at release, by automated tools and by hand, keyboard and screen reader included.
Go deeper
Contact us to validate the technical feasibility,
business impact and build your business case together.