Use of Artificial Intelligence

Last updated: July 21, 2026

Artificial intelligence is part of how we work at Mencova Tech, both in our own development process and in the products we build for clients: AI agents, integrations with language models (LLMs), and retrieval-augmented generation (RAG) architectures. This page explains how we use it and what controls we apply.

1. AI in our internal process

We use AI tools (code assistants, documentation generation, data analysis) to speed up repetitive tasks in our work. All code and deliverables go through human review by our team before reaching production or the client: AI speeds up the work, it doesn't replace it.

2. AI in the software we build

When a project includes AI agents or RAG architectures, this is explicitly defined in the project scope. In these cases:

• The client knows which AI model or provider is being used (for example, OpenAI, Anthropic, Google) and under what conditions.
• We design systems with human oversight at the points that require it, especially for sensitive business decisions.
• RAG architectures are built to answer based on the client's real information, not to generate unsourced, made-up content.

3. Client data and AI models

We do not use a client's confidential data to train or improve third-party or our own AI models, unless the client expressly authorizes it in writing. When a project connects to an external AI provider (for example, for an agent or a RAG system), that data is processed under that provider's privacy and retention policies, which we share with the client before implementing the solution.

4. Limitations

AI systems can make mistakes or generate incorrect responses ("hallucinations"). That's why we design the agents and RAG systems we build with verification mechanisms, clear scope limits, and, where applicable, the ability for a person to review or approve critical actions.

5. Contact

If you have questions about how we use AI in a specific project, write to us at info@mencovatech.com.