AI Disclosure

• What we do: AI helps you answer, write, translate and summarize.
• Your data is used to help you, and nothing else: NeoAgora does not use it to train models.
• In written conversations and tickets, detected personal information is masked before it is sent to the AI.
• Our databases are in Canada; only the text needed for a task is sent to the AI.
• A person stays in control of what is published on your behalf.

The platform uses recognized AI services for certain features: your website's conversational agents, portal assistants, writing help (articles, posts, newsletters), translation, summaries, images, synthetic voices and answers to support tickets.

An AI feature receives only the text needed for the task: for example the question asked and the relevant excerpts of your knowledge base. Your database is never sent in bulk.

In written conversations (your website's agent, the portal assistant, the intranet assistants), in support ticket answers and in newsletters, the personal information we detect is masked before it is sent to the AI model. It is put back in the answer you receive.

Text sent to an AI service may be processed outside Quebec and Canada, over an encrypted connection, under that provider's terms.

Your data is used only to provide the service you asked for: we do not sell it, do not use it for advertising and do not use it to train AI models. Our conversational agents, assistants, ticket answers, the writing of articles, posts and newsletters, and translation go through a professional AI service whose contract excludes using your texts to train models. If that service is down, a backup service may take over, with personal information already masked.

Our main providers hold recognized certifications, verified in October 2026:
• data centres hosting our servers: ISO/IEC 27001;
• intranet database: SOC 2 Type II and ISO/IEC 27001;
• off-site backups: SOC 2 Type II;
• web application firewall and network protection: SOC 2 Type II, ISO/IEC 27001, 27018 and 27701;
• main artificial intelligence service: SOC 2, ISO/IEC 27001 and ISO/IEC 42001 (AI management);
• payment processing: PCI DSS and SOC 2.

We keep a register of the providers that receive personal information: service provided, information received, place of processing and applicable agreement.

Our databases and your files are hosted in Canada, on redundant infrastructure with a standby copy kept up to date continuously. It is backed up several times a day; backups are encrypted and stored in Canada, and restoring them is tested.

All connections are encrypted. Each organization's data is kept separate from other clients' data, and each person sees only what their role allows.

AI-written content (article, post, newsletter) is first offered to you: you review, approve or schedule it. It goes out without review only if you turn on automatic publishing yourself; you then remain responsible for it.

A conversational agent's answers are produced live, without prior review: the following sections describe their limits.

Conversations with agents are kept in our databases in Canada for security, quality monitoring, abuse prevention and compliance. They are accessible to authorized NeoAgora staff and to the administrators of the client's account.

An agent's answers are for information only. They are not legal, accounting, tax, medical, psychological, financial or any other professional advice. A decision with significant consequences must be validated by a qualified person.

AI can confidently produce content that is inaccurate, outdated, incomplete or made up ("hallucinations"): check every fact, figure and citation before relying on it.

Do not give an agent sensitive information that is not needed.

You have the right to know that you are talking to an AI and to ask a person to review a significant decision about you. No individual decision with legal effects is made by AI alone. An organization running an agent must offer a clear way to reach a person.

NeoAgora is not responsible for decisions made solely on the basis of an agent's answers. The client configures its agents for its context and makes sure, before publishing generated content, that it complies with the law and the rights of others.

A person in charge of the protection of personal information has been designated and can be reached at [email protected].

In the event of a confidentiality incident, we contain it, assess the risk of serious injury and, where that risk exists, promptly notify the Commission d'accès à l'information and the persons concerned. The affected client organization is informed without delay. Every incident is recorded in our register.

Since October 5, 2026, NeoAgora has been listed on the Cloud Security Alliance public registry at Level 1 of CSA STAR and STAR for AI (self-assessments based on recognized industry questionnaires).

AI models evolve quickly: we may change model versions or providers, which may change how agents behave. The commitments in this notice stay the same, and significant changes are announced through the usual channels. Questions: [email protected].