Security

Last updated: September 30, 2026


Security is not a feature we bolt on. It is a posture we hold. This page describes, in plain language, how we approach protecting your data.

Our security principles

  1. Your source of truth stays yours. Your conversations and documents stay in your organization's own systems. We keep only the working memory Margot derives from them, and erase it when you leave.
  2. Margot works through her own accounts. She is connected by your administrator, works with the access your organization grants, and never signs in as any of your people.
  3. The platform decides access, not the AI. What Margot may see and do is determined by the platform before any AI is involved.
  4. No data sale, no AI training on your content.

How we protect your data

We use administrative, technical, and physical safeguards appropriate to the sensitivity of the information we hold — including encryption, separation between customer organizations, and restricted, role-based access for our staff. We run on established cloud infrastructure and keep our systems up to date.

Within your boundaries

Margot reaches only the tools your organization connects and the information it gives her access to. Your organization can narrow or remove that access at any time.

How AI accesses your data

When Margot needs information, the platform retrieves it through ordinary application code and decides what she may see. The platform is designed so that the AI cannot expand its own access.

The AI thinks; the platform decides what it gets to think about.

Service providers

We rely on trusted providers to host and operate Hey Margot, under written agreements that require them to protect your information.

Reporting a vulnerability

If you believe you have found a security issue, write to info@heymargot.ca with the subject line "Security report". We will work with you in good faith to validate and address it.

Incidents

If a confidentiality incident presents a risk of serious harm, we will act to contain it and notify the people affected and the authorities, as the law requires.

AI safety

Our acknowledgment

Perfect security does not exist. We will not pretend otherwise. We commit to holding ourselves to these principles, telling you the truth when something goes wrong, and improving continuously.