AI Policy
How we use AI, its limitations, and human oversight.
Draft for legal review. This is an initial English AI policy. It must be reviewed and adapted before publication.
AI in our services
We implement AI solutions such as RAG, fine-tuning, agents, and automation. We also plan to launch MarIA, an AI assistant for quotes and customer questions.
Limitations
AI can make mistakes, omit information, or produce unexpected results. Outputs require human review, especially for prices, schedules, legal, financial, or technical decisions. We do not guarantee accuracy or completeness.
Human oversight
Formal quotes and commitments require human approval. MarIA may collect requirements and prepare a draft, but a person reviews and confirms the final quote and terms.
Data
We avoid sending sensitive personal data to AI providers. We use minimization, access controls, and contractual safeguards where applicable. Client project data is governed by the service contract and DPA.
Client AI projects
For AI projects, we define data ownership, permitted use, evaluation, security, retention, and subprocessors before starting. We do not train models on client data for other clients without authorization.
Contact
Questions about AI: jose@tecnicosya.com.