Human judgment first
Delibia prepares the work and people decide. An answer from the tool is a starting point: an employee reviews it and corrects it if needed before using it.
Responsible AI policy
Responsible AI policy
Delibia helps municipal teams find information and prepare their documents. This policy describes the principles that guide the design of the tool and the commitments we make to municipalities.
Our principles
Sourced answers
Every answer cites the documents it relies on, with a link to the original, so you can check it.
Transparency about limits
We publish the known limits of the tool. The sources cited in each answer let you spot an error.
Known limits
An AI can be wrong. That is why every answer cites its sources: check them before any official use.
What we will never do
- Train a model on your documents or your queries.
- Sell or share your data.
- Decide in place of an elected official or an employee.
- Publish a document without human review.
- Let the model provider retain your queries.
Reference frameworks
- Law 25 (An Act to modernize legislative provisions as regards the protection of personal information).
- The Government of Québec's policy statement on digital sovereignty and information-technology procurement (February 2026).