Security and data handling
AI systems touch sensitive data and can take real actions. These are the principles we apply on every project. They describe how we work, not a certification.
Your data stays with you
Wherever possible, systems run in your own cloud accounts and read from your own data stores. We do not keep copies of your data on our machines beyond what a task strictly requires, and we delete working copies at the end of an engagement.
Least-privilege access
Every agent, workflow and connector gets the narrowest access it needs: specific systems, specific records, read-only by default. Write actions are added one at a time and reviewed with you.
No training on your data
We do not use client data to train or fine-tune models for anyone else. We configure model providers through business or API terms that exclude customer data from training.
People in the loop
Each action can run automatically, require approval or be sampled for review. Sensitive actions, such as sending messages to customers or changing records, can always require a person to confirm.
GDPR and LGPD awareness
We design with the EU General Data Protection Regulation and Brazil’s Lei Geral de Proteção de Dados in mind: data minimisation, clear purposes, data-location choices and support for your data-processing agreements. We are not lawyers, and we work with your legal or privacy team where needed.
Your choice of model provider
We are not tied to a single AI vendor. We recommend providers based on your data, location and compliance needs, and we build systems so the model can be changed later.
Logging and audit trails
Every agent decision, tool call and approval is logged with enough context to answer "what happened and why" weeks later. Logs live in your environment and follow your retention rules.
What we do not claim
We do not currently hold formal security certifications. We will happily complete your security questionnaire, work within your vendor-assessment process and sign a data-processing agreement.
Tell us what is taking up your team’s time.
A short message is enough. We reply with an honest view of what AI can and cannot do for that problem.
Start a conversation